What the Apple M3 Ultra (96 GB) can run

We check your GPU or device against every model. Browsers can’t see how much memory you have, so confirm your exact model below.

Unified memory
96.0 GB
Usable for models
72.0 GB
Memory bandwidth
819 GB/s
FP16 compute
28.4 TFLOPS
528 / 622 LLMs run Best: Gemma 4 26B-A4B (A+) 61 / 61 Image models run Best: Krea 2 Turbo (A+) 43 / 43 Video models run Best: LTX-2.5 (22B, video + audio) (A+) 70 / 70 Decision models run Best: Jina Reranker v3.5 (A+)

Open in Model Finder Best LLMs by task Compare with another GPU

LLMs on the Apple M3 Ultra (96 GB)

Quality Model Quant Memory Speed Runs
92Excellent quality (A+) · 92 of 100: how capable the model is at F16, whatever it runs on. Match score 96/100 (quality, fit, speed and memory use) sets the ranking.Gemma 4 26B-A4BGoogle · 26B (3.8B active)F1658.4 GB43.3 tok/sRuns great
93Excellent quality (A+) · 93 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 96/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.6 35B-A3BAlibaba · 36B (3.0B active)Q8_044.6 GB94.2 tok/sRuns great
93Excellent quality (A+) · 93 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 96/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.5 35B-A3BAlibaba · 36B (3.0B active)Q8_044.6 GB94.2 tok/sRuns great
93Excellent quality (A+) · 93 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 96/100 (quality, fit, speed and memory use) sets the ranking.MiniCPM-V 4.7 35B-A3BOpenBMB · 35B (3.0B active)Q8_043.9 GB94.2 tok/sRuns great
90Excellent quality (A+) · 90 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 95/100 (quality, fit, speed and memory use) sets the ranking.Flint (40B-A2.5B)Recursive · 40B (2.5B active)Q8_048.8 GB106 tok/sRuns great
92Excellent quality (A+) · 92 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 95/100 (quality, fit, speed and memory use) sets the ranking.Nex N2 Mini (35B-A3B)Nex-AGI · 35B (3.0B active)Q8_043.7 GB94.2 tok/sRuns great
89Excellent quality (A+) · 89 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 95/100 (quality, fit, speed and memory use) sets the ranking.Nemotron 3.5 Bolt 74B-A9BNVIDIA · 75B (9.0B active)Q5_K_M59.6 GB62.2 tok/sRuns great
91Excellent quality (A+) · 91 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 95/100 (quality, fit, speed and memory use) sets the ranking.Sage-A1.5 35B-A3BInternScience · 36B (3.0B active)Q8_044.6 GB94.2 tok/sRuns great
91Excellent quality (A+) · 91 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 94/100 (quality, fit, speed and memory use) sets the ranking.KAT-Coder V2.5 Dev (35B-A3B)Kuaishou · 35B (3.0B active)Q8_043.3 GB94.2 tok/sRuns great
90Excellent quality (A+) · 90 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 94/100 (quality, fit, speed and memory use) sets the ranking.Ornith 1.5 35B-A3BDeepReinforce · 36B (3.0B active)Q8_044.6 GB94.2 tok/sRuns great
90Excellent quality (A+) · 90 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 94/100 (quality, fit, speed and memory use) sets the ranking.Kolibri 1 (78B-A3.5B)Aleph Alpha · 78B (3.5B active)Q6_K70.6 GB107 tok/sRuns welltight fit
89Excellent quality (A+) · 89 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 94/100 (quality, fit, speed and memory use) sets the ranking.Apodex 1.1 Mini (35B-A3B)Apodex · 36B (3.0B active)Q8_044.6 GB94.2 tok/sRuns great
89Excellent quality (A+) · 89 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 94/100 (quality, fit, speed and memory use) sets the ranking.Ornith 1.0 35B-A3BDeepReinforce · 36B (3.0B active)Q8_044.6 GB94.2 tok/sRuns great
87Excellent quality (A+) · 87 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 93/100 (quality, fit, speed and memory use) sets the ranking.Agents-A1 35B-A3BInternScience · 35B (3.0B active)Q8_043.7 GB94.2 tok/sRuns great
87Excellent quality (A+) · 87 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 93/100 (quality, fit, speed and memory use) sets the ranking.Qwen AgentWorld 35B-A3BAlibaba · 35B (3.0B active)Q8_043.3 GB94.2 tok/sRuns great
88Excellent quality (A+) · 88 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 92/100 (quality, fit, speed and memory use) sets the ranking.GLM-4.7 Flash (30B-A3B)Zhipu AI · 31B (3.0B active)Q8_039.9 GB94.2 tok/sRuns great
85Excellent quality (A+) · 85 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 92/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 Coder Next (80B-A3B)Alibaba · 80B (3.0B active)Q6_K71.9 GB116 tok/sRuns welltight fit
81Great quality (A) · 81 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking.Qwen3-Next 80B-A3B ThinkingAlibaba · 81B (3.0B active)Q4_K_M56.2 GB146 tok/sRuns great
88Excellent quality (A+) · 88 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking.Raon-1 mini Preview (30B-A4B)KRAFTON · 29B (4.3B active)Q8_036.9 GB73.7 tok/sRuns great
79Great quality (A) · 79 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 90/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 Next 80B-A3B InstructAlibaba · 81B (3.0B active)Q4_K_M56.2 GB146 tok/sRuns great
85Excellent quality (A+) · 85 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 90/100 (quality, fit, speed and memory use) sets the ranking.Hunyuan Hy-Memory 30B-A3BTencent · 30B (3.0B active)Q8_039.0 GB94.2 tok/sRuns great
82Great quality (A) · 82 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 90/100 (quality, fit, speed and memory use) sets the ranking.Laguna XS 2.1 (33B-A3B)Poolside · 33B (3.0B active)Q8_043.1 GB94.2 tok/sRuns great
78Great quality (A) · 78 of 100: how capable the model is at F16, whatever it runs on. Match score 90/100 (quality, fit, speed and memory use) sets the ranking.DiffusionGemma 26B-A4B ITGoogle · 26B (4.0B active)F1658.4 GB41.5 tok/sRuns great
78Great quality (A) · 78 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 89/100 (quality, fit, speed and memory use) sets the ranking.Nemotron 3.5 Super VL 120B-A12BNVIDIA · 124B (12.0B active)Q3_K_M67.4 GB71.6 tok/sRuns welltight fit
77Great quality (A) · 77 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 89/100 (quality, fit, speed and memory use) sets the ranking.Ornith 1.5 122B-A10BDeepReinforce · 125B (10.0B active)Q3_K_M67.9 GB82.1 tok/sRuns welltight fit
80Great quality (A) · 80 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking.Nemotron Cascade 2 30B-A3BNVIDIA · 32B (3.5B active)Q8_039.9 GB85.1 tok/sRuns great
77Great quality (A) · 77 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.5 122B-A10BAlibaba · 125B (10.0B active)Q3_K_M67.9 GB82.1 tok/sRuns welltight fit
80Great quality (A) · 80 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking.North Mini Code 1.0 (30B-A3B)Cohere · 30B (3.0B active)Q8_039.0 GB94.2 tok/sRuns great
74Great quality (A) · 74 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking.Kimi Linear 48B-A3BMoonshot AI · 49B (3.0B active)Q8_058.5 GB94.2 tok/sRuns great
74Great quality (A) · 74 of 100: how capable the model is at F16, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking.LFM2.5 24B-A2BLiquid AI · 24B (2.5B active)F1654.5 GB60.8 tok/sRuns great
76Great quality (A) · 76 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking.Ling 2.6 Flash (104B-A7.4B)Ant Group (inclusionAI) · 107B (7.4B active)Q4_K_M72.0 GB83.9 tok/sRuns welltight fit
76Great quality (A) · 76 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking.Solar Open 100BUpstage · 103B (12.0B active)Q4_K_M70.6 GB58.0 tok/sRuns welltight fit
78Great quality (A) · 78 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 30B-A3B Thinking 2507Alibaba · 31B (3.3B active)Q8_039.5 GB88.6 tok/sRuns great
78Great quality (A) · 78 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 Coder 30B-A3BAlibaba · 31B (3.3B active)Q8_039.5 GB88.6 tok/sRuns great
77Great quality (A) · 77 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.Nemotron 3 Nano 30B-A3BNVIDIA · 32B (3.5B active)Q8_039.9 GB85.1 tok/sRuns great
74Great quality (A) · 74 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.Ring Flash 2.0 (100B-A6B)Ant Group (inclusionAI) · 103B (6.1B active)Q4_K_M69.7 GB96.0 tok/sRuns welltight fit
77Great quality (A) · 77 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 30B-A3B Instruct 2507Alibaba · 31B (3.3B active)Q8_039.5 GB88.6 tok/sRuns great
73Great quality (A) · 73 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.Nemotron 3 Super 120B-A12BNVIDIA · 124B (12.0B active)Q3_K_M67.0 GB71.6 tok/sRuns welltight fit
90Excellent quality (A+) · 90 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.Nanbeige4.3 33BBOSS Zhipin (Nanbeige) · 33BQ4_K_M28.3 GB26.1 tok/sRuns well
76Great quality (A) · 76 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.Qwen3-VL 30B-A3B ThinkingAlibaba · 31B (3.3B active)Q8_040.1 GB88.6 tok/sRuns great
90Excellent quality (A+) · 90 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.8 27BAlibaba · 28BQ4_K_M23.8 GB31.2 tok/sRuns great
70Great quality (A) · 70 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.Hunyuan A13B (80B-A13B)Tencent · 80B (13.0B active)Q4_K_M56.5 GB54.4 tok/sRuns great
72Great quality (A) · 72 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.Laguna S 2.1 (118B-A8B)Poolside · 118B (8.0B active)Q3_K_M65.6 GB96.3 tok/sRuns welltight fit
70Great quality (A) · 70 of 100: how capable the model is at F16, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.ERNIE 4.5 21B-A3B ThinkingBaidu · 22B (3.0B active)F1650.5 GB52.2 tok/sRuns great
70Great quality (A) · 70 of 100: how capable the model is at F16, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.gpt-oss 20BOpenAI · 21B (3.6B active)F1648.4 GB45.3 tok/sRuns great
70Great quality (A) · 70 of 100: how capable the model is at F16, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.Maple Preview (20B-A1B)DeepGrove · 20B (1.5B active)F1646.9 GB85.4 tok/sRuns great
72Great quality (A) · 72 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.Ling Flash 2.0 (100B-A6B)Ant Group (inclusionAI) · 103B (6.1B active)Q4_K_M69.7 GB96.0 tok/sRuns welltight fit
69Good quality (B) · 69 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.GLM-4.6VZhipu AI · 108B (12.0B active)Q3_K_M60.7 GB71.6 tok/sRuns great
74Great quality (A) · 74 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.Qwen3-VL 30B-A3B InstructAlibaba · 31B (3.3B active)Q8_040.1 GB88.6 tok/sRuns great
89Excellent quality (A+) · 89 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.6 27BAlibaba · 28BQ4_K_M23.8 GB31.2 tok/sRuns great
71Great quality (A) · 71 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.Sarvam 105BSarvam AI · 106B (10.0B active)Q4_K_M71.3 GB67.0 tok/sRuns welltight fit
74Great quality (A) · 74 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 30B-A3BAlibaba · 31B (3.3B active)Q8_039.5 GB88.6 tok/sRuns great
70Great quality (A) · 70 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.Qwen3-Omni 30B-A3B InstructAlibaba · 35B (3.3B active)Q8_044.5 GB88.6 tok/sRuns great
88Excellent quality (A+) · 88 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.5 27BAlibaba · 28BQ4_K_M23.8 GB31.2 tok/sRuns great
67Good quality (B) · 67 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.GLM-4.5VZhipu AI · 108B (12.0B active)Q3_K_M60.7 GB71.6 tok/sRuns great
72Great quality (A) · 72 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.Nemotron 3.5 Lightning 30B-A3BNVIDIA · 32B (3.0B active)Q8_039.9 GB94.2 tok/sRuns great
86Excellent quality (A+) · 86 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.Gemma 4 31BGoogle · 31BQ4_K_M27.0 GB27.7 tok/sRuns well
87Excellent quality (A+) · 87 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.Muse Glimmer 30BMeta · 30BQ4_K_M24.6 GB29.1 tok/sRuns well
68Good quality (B) · 68 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking.Ling 3.0 Flash (127B-A5B)Ant Group (inclusionAI) · 127B (5.1B active)Q3_K_M69.2 GB128 tok/sRuns welltight fit
66Good quality (B) · 66 of 100: how capable the model is at F16, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking.Trinity Mini (26B-A3B)Arcee AI · 26B (3.0B active)F1658.8 GB52.2 tok/sRuns great
70Great quality (A) · 70 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking.Sarvam 30BSarvam AI · 32B (2.4B active)Q8_040.6 GB108 tok/sRuns great
67Good quality (B) · 67 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking.Mistral Small 4 (119B-A6B)Mistral AI · 119B (6.0B active)Q3_K_M65.1 GB116 tok/sRuns welltight fit
85Excellent quality (A+) · 85 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking.TMax 27BAi2 · 27BQ4_K_M23.6 GB31.6 tok/sRuns great
64Good quality (B) · 64 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.ERNIE 4.5 21B-A3BBaidu · 22B (3.0B active)F1650.7 GB52.2 tok/sRuns great
83Great quality (A) · 83 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.Fara 1.5 27BMicrosoft · 27BQ4_K_M23.6 GB31.6 tok/sRuns great
65Good quality (B) · 65 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.GLM-4.5 Air (106B-A12B)Zhipu AI · 110B (12.0B active)Q3_K_M62.0 GB71.6 tok/sRuns welltight fit
81Great quality (A) · 81 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.UI-Mate 27BTencent · 27BQ4_K_M23.6 GB31.6 tok/sRuns great
76Great quality (A) · 76 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.Kimi Dev 72BMoonshot AI · 73BQ4_K_M53.4 GB11.9 tok/sRuns well
82Great quality (A) · 82 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.Gemma 4 12BGoogle · 12BQ6_K16.8 GB50.2 tok/sRuns great
60Good quality (B) · 60 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.Llama 4 Scout 17B-16EMeta · 109B (17.0B active)Q3_K_M61.2 GB54.3 tok/sRuns great
62Good quality (B) · 62 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.Leanstral 119B-A6BMistral AI · 119B (6.5B active)Q3_K_M65.1 GB111 tok/sRuns welltight fit
62Good quality (B) · 62 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.Axion 1.0 35BAxion · 35B (3.0B active)Q8_043.7 GB94.2 tok/sRuns great
76Great quality (A) · 76 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.3 Nemotron Super 49B v1.5NVIDIA · 50BQ4_K_M39.5 GB17.4 tok/sRuns well
80Great quality (A) · 80 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.K2 Horizon 7BInstitute of Foundation Models · 9.0BQ8_017.1 GB49.0 tok/sRuns great
70Great quality (A) · 70 of 100: how capable the model is at F16, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.Mellum2.1 Thinking (12B-A2.5B)JetBrains · 12B (2.5B active)F1630.8 GB59.9 tok/sRuns great
77Great quality (A) · 77 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.EXAONE 4.5 33BLG AI Research · 34BQ4_K_M28.1 GB25.2 tok/sRuns well
62Good quality (B) · 62 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.Granite 4.0 H Small (32B-A9B)IBM · 32B (9.0B active)Q8_040.6 GB41.3 tok/sRuns great
76Great quality (A) · 76 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.Seed-OSS 36B InstructByteDance · 36BQ4_K_M30.6 GB23.9 tok/sRuns well
78Great quality (A) · 78 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.SenseNova Loop 9B PreviewSenseTime · 9.7BQ8_017.1 GB45.6 tok/sRuns great
59Good quality (B) · 59 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.NorthCUA Minibeipo-ai · 35B (3.0B active)Q8_043.7 GB94.2 tok/sRuns great
58Good quality (B) · 58 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.Groot 65B A6B 261023GrootTeam · 118B (6.0B active)Q3_K_M64.5 GB116 tok/sRuns welltight fit
78Great quality (A) · 78 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.MiMo V2.6 Distill Qwen 9BXiaomi · 9.4BQ8_016.6 GB46.8 tok/sRuns great
56Good quality (B) · 56 of 100: how capable the model is at F16, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.LFM2 24B-A2BLiquid AI · 24B (2.3B active)F1654.2 GB63.7 tok/sRuns great
56Good quality (B) · 56 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.Phi-3.5 MoE (42B-A6.6B)Microsoft · 42B (6.6B active)Q8_051.8 GB53.2 tok/sRuns great
57Good quality (B) · 57 of 100: how capable the model is at F16, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.Groot 19B A3B 261023GrootTeam · 19B (3.0B active)F1645.4 GB52.2 tok/sRuns great
75Great quality (A) · 75 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.Hermes 4.3 36BNous Research · 36BQ4_K_M30.6 GB23.9 tok/sRuns well
77Great quality (A) · 77 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.Ornith 1.5 9BDeepReinforce · 9.7BQ8_016.9 GB45.7 tok/sRuns great
70Great quality (A) · 70 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.Hermes 4 70BNous Research · 71BQ4_K_M52.1 GB12.3 tok/sRuns well
70Great quality (A) · 70 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.3 70B InstructMeta · 71BQ4_K_M52.1 GB12.3 tok/sRuns well
57Good quality (B) · 57 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.Ling 3.0 flash SanteAnt Group (inclusionAI) · 127B (5.1B active)Q3_K_M68.9 GB128 tok/sRuns welltight fit
74Great quality (A) · 74 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.Qwen3-VL 32B InstructAlibaba · 33BQ4_K_M28.9 GB26.0 tok/sRuns well
74Great quality (A) · 74 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.QwQ 32BAlibaba · 33BQ4_K_M28.5 GB26.4 tok/sRuns well
60Good quality (B) · 60 of 100: how capable the model is at F16, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.Kimi-VL A3B Thinking 2506Moonshot AI · 16B (3.0B active)F1639.4 GB52.2 tok/sRuns great
76Great quality (A) · 76 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.5 9BAlibaba · 9.7BQ8_016.9 GB45.7 tok/sRuns great
70Great quality (A) · 70 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.InternVL3 78BShanghai AI Lab · 78BQ4_K_M56.9 GB11.0 tok/sRuns well
69Good quality (B) · 69 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 72B InstructAlibaba · 73BQ4_K_M53.4 GB11.9 tok/sRuns well
69Good quality (B) · 69 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5-VL 72B InstructAlibaba · 73BQ4_K_M53.8 GB11.8 tok/sRuns well
75Great quality (A) · 75 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.Ornith 1.0 9BDeepReinforce · 9.7BQ8_016.9 GB45.7 tok/sRuns great
68Good quality (B) · 68 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek R1 Distill Llama 70BDeepSeek · 71BQ4_K_M52.1 GB12.3 tok/sRuns well
68Good quality (B) · 68 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.Tülu 3 70BAi2 · 71BQ4_K_M52.1 GB12.3 tok/sRuns well
72Great quality (A) · 72 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.InternVL3.5 38BShanghai AI Lab · 38BQ4_K_M31.9 GB22.6 tok/sRuns well
72Great quality (A) · 72 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.Devstral Small 2 (24B)Mistral AI · 24BQ4_K_M22.3 GB36.1 tok/sRuns great
72Great quality (A) · 72 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.Granite 4.2 30BIBM · 29BQ4_K_M26.4 GB29.6 tok/sRuns well
72Great quality (A) · 72 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 32BAlibaba · 33BQ4_K_M28.5 GB26.4 tok/sRuns well
72Great quality (A) · 72 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.OpenReasoning Nemotron 32BNVIDIA · 33BQ4_K_M28.5 GB26.4 tok/sRuns well
72Great quality (A) · 72 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.Skywork OR1 32BSkywork · 33BQ4_K_M28.5 GB26.4 tok/sRuns well
70Great quality (A) · 70 of 100: how capable the model is at F16, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.Nemotron 3.5 Bolt E4BNVIDIA · 8.0BF1622.5 GB45.1 tok/sRuns great
71Great quality (A) · 71 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.Apriel 1.5 15B ThinkerServiceNow · 15BQ6_K20.2 GB40.4 tok/sRuns great
71Great quality (A) · 71 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.Phi-4 Reasoning Plus (14B)Microsoft · 15BQ6_K20.1 GB41.0 tok/sRuns great
74Great quality (A) · 74 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.TMax 9BAi2 · 9.0BQ8_016.1 GB49.2 tok/sRuns great
72Great quality (A) · 72 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.Olmo 3.1 32B ThinkAi2 · 32BQ4_K_M26.8 GB26.9 tok/sRuns well
66Good quality (B) · 66 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.1 70B InstructMeta · 71BQ4_K_M52.1 GB12.3 tok/sRuns well
70Great quality (A) · 70 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.EXAONE Deep 32BLG AI Research · 32BQ4_K_M28.0 GB27.1 tok/sRuns well
53Fair quality (C) · 53 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.dots.llm1 Instructrednote · 143B (14.0B active)Q2_K71.8 GB77.1 tok/sRuns welltight fit
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Gemma 4 E4BGoogle · 8.0BF1622.5 GB45.1 tok/sRuns great
84Great quality (A) · 84 of 100: how capable the model is at MXFP4, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.gpt-oss 120BOpenAI · 117B (5.1B active)MXFP473.3 GB95.8 tok/sRuns wellover GPU memory limit
51Fair quality (C) · 51 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Nous Hermes 2 Mixtral 8x7B DPONous Research · 47B (12.9B active)Q6_K45.7 GB39.9 tok/sRuns great
69Good quality (B) · 69 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Phi-4 Reasoning Vision 15BMicrosoft · 15BQ6_K20.4 GB39.7 tok/sRuns great
69Good quality (B) · 69 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Phi-4 Reasoning (14B)Microsoft · 15BQ6_K20.1 GB41.0 tok/sRuns great
68Good quality (B) · 68 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Command A Reasoning (111B)Cohere · 111BQ3_K_M61.5 GB10.1 tok/sRuns welltight fit
70Great quality (A) · 70 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.GLM-4 32B 0414Zhipu AI · 33BQ4_K_M26.7 GB26.6 tok/sRuns well
70Great quality (A) · 70 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.GLM-Z1 32B (0414)Zhipu AI · 33BQ4_K_M26.7 GB26.6 tok/sRuns well
64Good quality (B) · 64 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Hermes 3 (Llama 3.1 70B)Nous Research · 71BQ4_K_M52.1 GB12.3 tok/sRuns well
49Fair quality (C) · 49 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Jamba 1.5 MiniAI21 Labs · 52B (12.0B active)Q6_K48.8 GB42.4 tok/sRuns great
64Good quality (B) · 64 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.QVQ 72B PreviewAlibaba · 73BQ4_K_M53.8 GB11.8 tok/sRuns well
68Good quality (B) · 68 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Magistral Small 1.2 (24B)Mistral AI · 24BQ4_K_M22.3 GB36.1 tok/sRuns great
68Good quality (B) · 68 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Granite 4.1 30BIBM · 29BQ4_K_M26.1 GB30.0 tok/sRuns well
68Good quality (B) · 68 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek R1 Distill Qwen 32BDeepSeek · 33BQ4_K_M28.5 GB26.4 tok/sRuns well
49Fair quality (C) · 49 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Mixtral 8x7B InstructMistral AI · 47B (12.9B active)Q6_K45.7 GB39.9 tok/sRuns great
48Fair quality (C) · 48 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Qwen2 57B-A14B InstructAlibaba · 57B (14.0B active)Q5_K_M47.6 GB43.3 tok/sRuns great
70Great quality (A) · 70 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.UI-Mate 9BTencent · 9.4BQ8_016.6 GB46.8 tok/sRuns great
70Great quality (A) · 70 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.UI-Venus 2 9BAnt Group (inclusionAI) · 9.4BQ8_016.6 GB46.8 tok/sRuns great
67Good quality (B) · 67 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Command A (111B)Cohere · 111BQ3_K_M61.5 GB10.1 tok/sRuns welltight fit
68Good quality (B) · 68 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Gemma 3 27B ITGoogle · 27BQ4_K_M23.9 GB31.6 tok/sRuns great
67Good quality (B) · 67 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 14BAlibaba · 15BQ6_K19.8 GB40.6 tok/sRuns great
68Good quality (B) · 68 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Olmo 3.1 32B InstructAi2 · 32BQ4_K_M26.8 GB26.9 tok/sRuns well
68Good quality (B) · 68 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.EXAONE 4.0 32BLG AI Research · 32BQ4_K_M26.6 GB27.1 tok/sRuns well
62Good quality (B) · 62 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.1 Nemotron 70B InstructNVIDIA · 71BQ4_K_M52.1 GB12.3 tok/sRuns well
62Good quality (B) · 62 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Apertus v1.5 70BSwiss AI · 72BQ4_K_M53.0 GB12.0 tok/sRuns well
66Good quality (B) · 66 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.InternVL3 38BShanghai AI Lab · 38BQ4_K_M31.9 GB22.6 tok/sRuns well
66Good quality (B) · 66 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Devstral Small 1.1 (24B)Mistral AI · 24BQ4_K_M22.1 GB36.7 tok/sRuns great
66Good quality (B) · 66 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Magistral Small 2507Mistral AI · 24BQ4_K_M22.1 GB36.7 tok/sRuns great
66Good quality (B) · 66 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 32B InstructAlibaba · 33BQ4_K_M28.5 GB26.4 tok/sRuns well
66Good quality (B) · 66 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Baichuan M2 32BBaichuan · 33BQ4_K_M28.5 GB26.4 tok/sRuns well
66Good quality (B) · 66 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.OpenHands LM 32BAll Hands AI · 33BQ4_K_M28.5 GB26.4 tok/sRuns well
68Good quality (B) · 68 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Granite 4.2 8BIBM · 8.8BQ8_017.0 GB50.1 tok/sRuns great
52Fair quality (C) · 52 of 100: how capable the model is at F16, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Ling Mini 2.0 (16B-A1.4B)Ant Group (inclusionAI) · 16B (1.4B active)F1639.2 GB88.7 tok/sRuns great
66Good quality (B) · 66 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.MedGemma 27B Text ITGoogle · 27BQ4_K_M23.6 GB32.1 tok/sRuns great
65Good quality (B) · 65 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Apriel Nemotron 15B ThinkerServiceNow · 15BQ6_K20.3 GB40.1 tok/sRuns great
65Good quality (B) · 65 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.AceReason Nemotron 14BNVIDIA · 15BQ6_K20.1 GB40.7 tok/sRuns great
65Good quality (B) · 65 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.OpenReasoning Nemotron 14BNVIDIA · 15BQ6_K20.1 GB40.7 tok/sRuns great
65Good quality (B) · 65 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 Coder 14B InstructAlibaba · 15BQ6_K20.1 GB40.7 tok/sRuns great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Nanbeige4.2 3BBOSS Zhipin (Nanbeige) · 4.2BF1616.2 GB48.7 tok/sRuns great
65Good quality (B) · 65 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Hermes 4 14BNous Research · 15BQ6_K19.8 GB40.7 tok/sRuns great
52Fair quality (C) · 52 of 100: how capable the model is at F16, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek Coder V2 Lite (16B)DeepSeek · 16B (2.4B active)F1638.0 GB61.8 tok/sRuns great
60Good quality (B) · 60 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Qwen2 72B InstructAlibaba · 73BQ4_K_M53.4 GB11.9 tok/sRuns well
65Good quality (B) · 65 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Ministral 3 14B ReasoningMistral AI · 14BQ6_K19.1 GB43.1 tok/sRuns great
64Good quality (B) · 64 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Magistral Small 2506Mistral AI · 24BQ4_K_M22.1 GB36.7 tok/sRuns great
60Good quality (B) · 60 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Qwen2-VL 72B InstructAlibaba · 73BQ4_K_M53.8 GB11.8 tok/sRuns well
64Good quality (B) · 64 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Mistral Small 3.2 24BMistral AI · 24BQ4_K_M22.3 GB36.1 tok/sRuns great
64Good quality (B) · 64 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5-VL 32B InstructAlibaba · 33BQ4_K_M28.9 GB25.9 tok/sRuns well
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Nemotron Labs Dory 1.0 4BNVIDIA · 4.0BF1614.8 GB51.0 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at F16, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Kimi-VL A3B InstructMoonshot AI · 16B (3.0B active)F1639.4 GB52.2 tok/sRuns great
60Good quality (B) · 60 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Devstral 2 (123B)Mistral AI · 125BQ2_K59.3 GB11.4 tok/sRuns well
63Good quality (B) · 63 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.InternVL3 14BShanghai AI Lab · 15BQ6_K20.4 GB39.7 tok/sRuns great
63Good quality (B) · 63 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 Coder 32B InstructAlibaba · 33BQ4_K_M28.5 GB26.4 tok/sRuns well
64Good quality (B) · 64 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Terminal RSI 27BTeriminal-RSI · 28BQ4_K_M23.8 GB31.2 tok/sRuns great
63Good quality (B) · 63 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Phi-4 (14B)Microsoft · 15BQ6_K20.1 GB40.9 tok/sRuns great
63Good quality (B) · 63 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.DeepCoder 14B PreviewAgentica · 15BQ6_K20.1 GB40.7 tok/sRuns great
62Good quality (B) · 62 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Reka Flash 3.1 (21B)Reka AI · 21BQ4_K_M20.2 GB41.4 tok/sRuns great
62Good quality (B) · 62 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Devstral Small 2505Mistral AI · 24BQ4_K_M22.1 GB36.7 tok/sRuns great
63Good quality (B) · 63 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Ministral 3 14BMistral AI · 14BQ6_K19.1 GB43.1 tok/sRuns great
62Good quality (B) · 62 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Mistral Small 3.1 24BMistral AI · 24BQ4_K_M22.3 GB36.1 tok/sRuns great
63Good quality (B) · 63 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Hemmingway 1.5Hemmingway · 27BQ4_K_M23.3 GB32.2 tok/sRuns great
62Good quality (B) · 62 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.QwQ 32B PreviewAlibaba · 33BQ4_K_M28.5 GB26.4 tok/sRuns well
66Good quality (B) · 66 of 100: how capable the model is at F16, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Spark-X2.6 4BSparkLLM · 4.1BF1614.9 GB49.4 tok/sRuns great
58Good quality (B) · 58 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Command R+ (08-2024) 104BCohere · 104BQ3_K_M59.4 GB10.8 tok/sRuns well
60Good quality (B) · 60 of 100: how capable the model is at F16, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Ling 3.0 Tiny (8B-A1.3B)Ant Group (inclusionAI) · 7.9B (1.3B active)F1622.4 GB92.8 tok/sRuns great
62Good quality (B) · 62 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.SenseNova 6.8 Code2VideoSenseTime · 27BQ4_K_M23.6 GB31.6 tok/sRuns great
61Good quality (B) · 61 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek R1 Distill Qwen 14BDeepSeek · 15BQ6_K20.1 GB40.7 tok/sRuns great
61Good quality (B) · 61 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 14B InstructAlibaba · 15BQ6_K20.0 GB40.9 tok/sRuns great
61Good quality (B) · 61 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.JanusCoder 14BShanghai AI Lab · 15BQ6_K19.8 GB40.7 tok/sRuns great
57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Apertus 70B InstructSwiss AI · 71BQ4_K_M52.1 GB12.3 tok/sRuns well
60Good quality (B) · 60 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Mistral Small 3 24B (2501)Mistral AI · 24BQ4_K_M22.1 GB36.7 tok/sRuns great
60Good quality (B) · 60 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Falcon-H1 34B InstructTII · 34BQ4_K_M28.1 GB25.7 tok/sRuns well
55Good quality (B) · 55 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.2 90B Vision InstructMeta · 89BQ3_K_M53.1 GB12.6 tok/sRuns well
62Good quality (B) · 62 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Qwen3-VL 8B InstructAlibaba · 8.8BQ8_016.9 GB50.3 tok/sRuns great
58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Codestral 22BMistral AI · 22BQ4_K_M21.8 GB38.9 tok/sRuns great
56Good quality (B) · 56 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Mistral Large 2.1 (2411)Mistral AI · 123BQ2_K58.4 GB11.6 tok/sRuns well
62Good quality (B) · 62 of 100: how capable the model is at F16, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Gemma 4 E2BGoogle · 5.1BF1616.7 GB88.3 tok/sRuns great
56Good quality (B) · 56 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Mistral Medium 3.5 (128B)Mistral AI · 128BQ2_K60.4 GB11.1 tok/sRuns well
62Good quality (B) · 62 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.InternVL3.5 8BShanghai AI Lab · 8.5BQ8_016.6 GB51.7 tok/sRuns great
59Good quality (B) · 59 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 14B Instruct 1MAlibaba · 15BQ6_K20.1 GB40.7 tok/sRuns great
62Good quality (B) · 62 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek R1 0528 Qwen3 8BDeepSeek · 8.2BQ8_016.2 GB53.8 tok/sRuns great
55Good quality (B) · 55 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Llama 3 70B InstructMeta · 71BQ4_K_M52.1 GB12.3 tok/sRuns well
46Fair quality (C) · 46 of 100: how capable the model is at F16, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Moonlight 16B-A3B InstructMoonshot AI · 16B (3.0B active)F1638.5 GB52.2 tok/sRuns great
62Good quality (B) · 62 of 100: how capable the model is at F16, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.5 4BAlibaba · 4.7BF1615.9 GB43.6 tok/sRuns great
55Good quality (B) · 55 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5-Math 72B InstructAlibaba · 73BQ4_K_M52.1 GB11.9 tok/sRuns well
58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Dolphin Mistral 24B Venice EditionCognitive Computations · 24BQ4_K_M22.3 GB36.1 tok/sRuns great
58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Voxtral Small 24BMistral AI · 24BQ4_K_M22.5 GB35.7 tok/sRuns great
58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Gemma 2 27B ITGoogle · 27BQ4_K_M24.7 GB31.8 tok/sRuns great
60Good quality (B) · 60 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.GLM-4.1V 9B ThinkingZhipu AI · 10BQ8_017.6 GB42.8 tok/sRuns great
54Fair quality (C) · 54 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Mistral Large 2 (123B)Mistral AI · 123BQ2_K58.5 GB11.6 tok/sRuns well
54Fair quality (C) · 54 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Pixtral Large 124B (2411)Mistral AI · 124BQ2_K58.9 GB11.5 tok/sRuns well
58Good quality (B) · 58 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Command R (08-2024) 32BCohere · 32BQ4_K_M27.4 GB26.8 tok/sRuns well
60Good quality (B) · 60 of 100: how capable the model is at F16, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Mage-VL-Cowork 4.7BMicrosoft · 4.7BF1617.0 GB42.9 tok/sRuns great
60Good quality (B) · 60 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Granite 4.1 8BIBM · 8.8BQ8_017.0 GB50.1 tok/sRuns great
60Good quality (B) · 60 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.MiniCPM-V 4.5 8BOpenBMB · 8.7BQ8_016.8 GB50.7 tok/sRuns great
57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Solar Pro Preview 22BUpstage · 22BQ4_K_M21.2 GB39.1 tok/sRuns great
61Good quality (B) · 61 of 100: how capable the model is at F16, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.TMax 4BAi2 · 4.2BF1615.0 GB48.2 tok/sRuns great
60Good quality (B) · 60 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 8BAlibaba · 8.2BQ8_016.2 GB53.8 tok/sRuns great
51Fair quality (C) · 51 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Command R 35B (v01)Cohere · 35BQ4_K_M38.4 GB24.7 tok/sRuns well
59Good quality (B) · 59 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Gemma 3 12B ITGoogle · 12BQ6_K17.0 GB49.2 tok/sRuns great
60Good quality (B) · 60 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Nemotron Nano 9B v2NVIDIA · 8.9BQ8_015.9 GB49.6 tok/sRuns great
57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Dolphin 3.0 R1 Mistral 24BCognitive Computations · 24BQ4_K_M22.1 GB36.7 tok/sRuns great
60Good quality (B) · 60 of 100: how capable the model is at F16, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Agents-A1 4BInternScience · 4.5BF1615.7 GB44.7 tok/sRuns great
53Fair quality (C) · 53 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.Command R+ 104B (04-2024)Cohere · 104BQ3_K_M59.4 GB10.8 tok/sRuns well
57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.Aya Vision 32BCohere · 33BQ4_K_M27.9 GB26.1 tok/sRuns well
57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.Aya Expanse 32BCohere · 32BQ4_K_M27.4 GB26.8 tok/sRuns well
57Good quality (B) · 57 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.OLMo 2 32B InstructAi2 · 32BQ4_K_M27.1 GB26.9 tok/sRuns well
55Good quality (B) · 55 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.LLM4Decompile 22B V2LLM4Binary · 22BQ4_K_M21.8 GB38.9 tok/sRuns great
58Good quality (B) · 58 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.GLM-Z1 9B (0414)Zhipu AI · 9.4BQ8_016.7 GB46.9 tok/sRuns great
58Good quality (B) · 58 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.MiMo-VL 7B RLXiaomi · 8.3BQ8_016.4 GB53.0 tok/sRuns great
58Good quality (B) · 58 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.Seed-Coder 8B ReasoningByteDance · 8.3BQ8_016.2 GB53.4 tok/sRuns great
58Good quality (B) · 58 of 100: how capable the model is at F16, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.Fara 1.5 4BMicrosoft · 4.5BF1615.7 GB44.7 tok/sRuns great
52Fair quality (C) · 52 of 100: how capable the model is at F16, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.LFM2.5 8B-A1BLiquid AI · 8.5B (1.5B active)F1623.4 GB85.0 tok/sRuns great
60Good quality (B) · 60 of 100: how capable the model is at F16, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.MiniCPM5 2BOpenBMB · 2.5BF1611.8 GB80.6 tok/sRuns great
56Good quality (B) · 56 of 100: how capable the model is at F16, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.Mage-VL 4.7BMicrosoft · 4.7BF1617.0 GB42.9 tok/sRuns great
56Good quality (B) · 56 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.Ministral 3 8BMistral AI · 8.9BQ8_017.0 GB49.4 tok/sRuns great
47Fair quality (C) · 47 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.Aya 23 35BCohere · 35BQ4_K_M38.4 GB24.7 tok/sRuns well
67Good quality (B) · 67 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.8 Flash Next (125B-A6B)Alibaba · 180B (6.0B active)Q2_K78.6 GB116 tok/sRuns wellover GPU memory limit
56Good quality (B) · 56 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.Seed-Coder 8B InstructByteDance · 8.3BQ8_016.2 GB53.4 tok/sRuns great
56Good quality (B) · 56 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.MiMo 7B RLXiaomi · 7.8BQ8_015.9 GB56.3 tok/sRuns great
56Good quality (B) · 56 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.EXAONE Deep 7.8BLG AI Research · 7.8BQ8_015.7 GB56.4 tok/sRuns great
56Good quality (B) · 56 of 100: how capable the model is at F16, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 4B Thinking 2507Alibaba · 4.0BF1615.6 GB50.5 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at F16, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V2 Lite ChatDeepSeek · 16B (2.4B active)F1638.0 GB61.8 tok/sRuns great
56Good quality (B) · 56 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.InternVL3 8BShanghai AI Lab · 7.9BQ8_015.2 GB55.5 tok/sRuns great
56Good quality (B) · 56 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.Falcon-H1R 7BTII · 7.6BQ8_014.8 GB58.1 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.Mixtral 8x22B InstructMistral AI · 141B (39.0B active)Q2_K64.5 GB32.7 tok/sRuns welltight fit
51Fair quality (C) · 51 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.Mistral Small 22B (2409)Mistral AI · 22BQ4_K_M21.8 GB38.9 tok/sRuns great
54Fair quality (C) · 54 of 100: how capable the model is at F16, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.Qwen3-VL 4B InstructAlibaba · 4.4BF1616.4 GB45.7 tok/sRuns great
55Good quality (B) · 55 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 Coder 7B InstructAlibaba · 7.6BQ8_014.9 GB57.8 tok/sRuns great
54Fair quality (C) · 54 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.1 Nemotron Nano 8B v1NVIDIA · 8.0BQ8_015.9 GB54.9 tok/sRuns great
52Fair quality (C) · 52 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.Falcon 3 10B InstructTII · 10BQ8_018.6 GB42.7 tok/sRuns great
54Fair quality (C) · 54 of 100: how capable the model is at F16, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 4B Instruct 2507Alibaba · 4.0BF1615.6 GB50.5 tok/sRuns great
54Fair quality (C) · 54 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.Olmo 3 7B ThinkAi2 · 7.3BQ8_015.6 GB60.4 tok/sRuns great
54Fair quality (C) · 54 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.MiniCPM4 8BOpenBMB · 8.2BQ8_015.3 GB53.8 tok/sRuns great
54Fair quality (C) · 54 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.OpenReasoning Nemotron 7BNVIDIA · 7.6BQ8_014.9 GB57.8 tok/sRuns great
51Fair quality (C) · 51 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.Mistral Nemo 12B InstructMistral AI · 12BQ6_K17.7 GB49.2 tok/sRuns great
52Fair quality (C) · 52 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.GLM-4 9B 0414Zhipu AI · 9.4BQ8_016.7 GB46.9 tok/sRuns great
52Fair quality (C) · 52 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.NorthCUA Nanobeipo-ai · 9.4BQ8_016.6 GB46.8 tok/sRuns great
52Fair quality (C) · 52 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.MiniCPM-o 2.6 8BOpenBMB · 8.7BQ8_016.0 GB50.8 tok/sRuns great
52Fair quality (C) · 52 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.Foundation-Sec 8B ReasoningCisco Foundation AI · 8.0BQ8_015.9 GB54.9 tok/sRuns great
52Fair quality (C) · 52 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5-VL 7B InstructAlibaba · 8.3BQ8_015.6 GB53.2 tok/sRuns great
56Good quality (B) · 56 of 100: how capable the model is at F16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.LFM3 1.7BLiquid AI · 1.7BF169.9 GB117 tok/sRuns great
52Fair quality (C) · 52 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.Hunyuan 7B InstructTencent · 7.5BQ8_015.4 GB58.8 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.Gemma 2 9B ITGoogle · 9.2BQ8_017.8 GB47.7 tok/sRuns great
52Fair quality (C) · 52 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 7B InstructAlibaba · 7.6BQ8_014.9 GB57.8 tok/sRuns great
49Fair quality (C) · 49 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.Pixtral 12B (2409)Mistral AI · 12BQ6_K17.9 GB48.4 tok/sRuns great
52Fair quality (C) · 52 of 100: how capable the model is at F16, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Granite 4.2 3BIBM · 3.7BF1614.4 GB55.5 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at F16, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Qianfan-OCR 4BBaidu · 4.7BF1617.0 GB42.9 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at F16, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Gemma 3n E4BGoogle · 7.9BF1622.2 GB50.8 tok/sRuns great
47Fair quality (C) · 47 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.InternLM2.5 20B ChatShanghai AI Lab · 20BQ4_K_M20.1 GB43.6 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.InternLM3 8B InstructShanghai AI Lab · 8.8BQ8_016.1 GB50.1 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.InternVL2.5 8BShanghai AI Lab · 8.1BQ8_016.0 GB54.5 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek R1 Distill Llama 8BDeepSeek · 8.0BQ8_015.9 GB54.9 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Foundation-Sec 1.1 8B InstructCisco Foundation AI · 8.0BQ8_015.9 GB54.9 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Fara 7BMicrosoft · 8.3BQ8_015.6 GB53.2 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at F16, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 4BAlibaba · 4.0BF1615.6 GB50.5 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Olmo 3 7B InstructAi2 · 7.3BQ8_015.6 GB60.4 tok/sRuns great
47Fair quality (C) · 47 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Yi 1.5 34B Chat01.AI · 34BQ4_K_M28.3 GB25.2 tok/sRuns well
50Fair quality (C) · 50 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.GLiDE 9BFastino · 7.9BQ8_015.0 GB55.5 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 7B Instruct 1MAlibaba · 7.6BQ8_014.9 GB57.8 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Falcon-H1 7B InstructTII · 7.6BQ8_014.8 GB58.1 tok/sRuns great
41Fair quality (C) · 41 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Code Llama 70B InstructMeta · 69BQ4_K_M49.8 GB12.6 tok/sRuns well
50Fair quality (C) · 50 of 100: how capable the model is at F16, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Nemotron 3 Nano 4BNVIDIA · 4.0BF1614.4 GB51.2 tok/sRuns great
48Fair quality (C) · 48 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.GLM-4 9B Chat 1MZhipu AI · 9.5BQ8_017.1 GB46.5 tok/sRuns great
48Fair quality (C) · 48 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.GLM-4 9B Chat (2024)Zhipu AI · 9.4BQ8_016.7 GB46.9 tok/sRuns great
45Fair quality (C) · 45 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.Phi-3 Medium 128K InstructMicrosoft · 14BQ6_K19.5 GB43.0 tok/sRuns great
48Fair quality (C) · 48 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.Hermes 3 (Llama 3.1 8B)Nous Research · 8.0BQ8_015.9 GB54.9 tok/sRuns great
48Fair quality (C) · 48 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.1 8B InstructMeta · 8.0BQ8_015.9 GB54.9 tok/sRuns great
48Fair quality (C) · 48 of 100: how capable the model is at F16, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.LightOnOCR-3 4BLightOn · 4.5BF1615.7 GB44.7 tok/sRuns great
48Fair quality (C) · 48 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.MiniCPM-V 2.6 8BOpenBMB · 8.1BQ8_015.4 GB54.4 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5-Omni 7BAlibaba · 11BQ8_018.2 GB41.1 tok/sRuns great
45Fair quality (C) · 45 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.Phi-3 Medium (14B)Microsoft · 14BQ6_K18.6 GB43.0 tok/sRuns great
43Fair quality (C) · 43 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.OLMo 2 13B InstructAi2 · 14BQ6_K21.0 GB43.8 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at F16, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.Granite 4.0 H Tiny (7B-A1B)IBM · 6.9B (1.0B active)F1620.3 GB108 tok/sRuns great
43Fair quality (C) · 43 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.Qwen1.5 110B ChatAlibaba · 111BQ3_K_M63.5 GB10.1 tok/sRuns welltight fit
46Fair quality (C) · 46 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.Apertus v1.5 8BSwiss AI · 8.9BQ8_016.9 GB49.5 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.CodeGeeX4 All 9BZhipu AI · 9.4BQ8_016.7 GB46.9 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Granite 3.2 8B InstructIBM · 8.2BQ8_016.4 GB53.9 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Granite 3.3 8BIBM · 8.2BQ8_016.4 GB53.9 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Ministral 8B (2410)Mistral AI · 8.0BQ8_016.1 GB55.0 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.DeepHermes 3 Llama 3 8B PreviewNous Research · 8.0BQ8_015.9 GB54.9 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Dolphin 3.0 Llama 3.1 8BCognitive Computations · 8.0BQ8_015.9 GB54.9 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Tülu 3 8BAi2 · 8.0BQ8_015.9 GB54.9 tok/sRuns great
50Fair quality (C) · 50 of 100: how capable the model is at F16, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.LFM3-VL 2BLiquid AI · 1.9BF1610.3 GB106 tok/sRuns great
45Fair quality (C) · 45 of 100: how capable the model is at F16, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.WeVisOne-VL 4BTencent · 4.8BF1617.2 GB42.1 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.EXAONE 3.5 7.8BLG AI Research · 7.8BQ8_015.7 GB56.4 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek LLM 67B ChatDeepSeek · 67BQ4_K_M49.1 GB12.8 tok/sRuns well
46Fair quality (C) · 46 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Falcon 3 7B InstructTII · 7.5BQ8_015.2 GB59.1 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at F16, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Phi-4 Mini Reasoning (3.8B)Microsoft · 3.8BF1615.1 GB52.9 tok/sRuns great
48Fair quality (C) · 48 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Chandra OCR 2Datalab · 5.3BQ8_012.2 GB83.2 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek R1 Distill Qwen 7BDeepSeek · 7.6BQ8_014.9 GB57.8 tok/sRuns great
46Fair quality (C) · 46 of 100: how capable the model is at F16, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Granite 4.1 3BIBM · 3.4BF1613.8 GB59.7 tok/sRuns great
41Fair quality (C) · 41 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Qwen1.5 32B ChatAlibaba · 33BQ4_K_M28.3 GB26.6 tok/sRuns well
45Fair quality (C) · 45 of 100: how capable the model is at F16, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Gemma 3 4B ITGoogle · 4.3BF1615.2 GB47.2 tok/sRuns great
41Fair quality (C) · 41 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek Coder 33B InstructDeepSeek · 33BQ4_K_M28.8 GB26.0 tok/sRuns well
44Fair quality (C) · 44 of 100: how capable the model is at F16, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.1 Nemotron Nano 4B v1.1NVIDIA · 4.5BF1616.5 GB45.0 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Granite 3.1 8B InstructIBM · 8.2BQ8_016.4 GB53.9 tok/sRuns great
42Fair quality (C) · 42 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.2 11B Vision InstructMeta · 11BQ8_019.0 GB41.3 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.InternVL2 8BShanghai AI Lab · 8.1BQ8_016.0 GB54.5 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Dolphin 2.9.4 Llama 3.1 8BCognitive Computations · 8.0BQ8_015.9 GB54.9 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Hermes 2 Pro Llama 3 8BNous Research · 8.0BQ8_015.9 GB54.9 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.OpenChat 3.6 8BOpenChat · 8.0BQ8_015.9 GB54.9 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Aya Vision 8BCohere · 8.6BQ8_015.9 GB51.1 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Qwen2-VL 7B InstructAlibaba · 8.3BQ8_015.6 GB53.2 tok/sRuns great
41Fair quality (C) · 41 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.LLaVA 1.6 34BLLaVA · 35BQ4_K_M28.6 GB24.9 tok/sRuns well
44Fair quality (C) · 44 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Phi-3 Small 8K InstructMicrosoft · 7.4BQ8_015.3 GB59.6 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Command R7BCohere · 8.0BQ8_015.2 GB54.9 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at F16, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.MedGemma 4B ITGoogle · 4.3BF1615.2 GB47.2 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at F16, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Phi-4 Mini (3.8B)Microsoft · 3.8BF1615.1 GB52.9 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at F16, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Ministral 3 3BMistral AI · 3.9BF1614.9 GB52.8 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Qwen2 7B InstructAlibaba · 7.6BQ8_014.9 GB57.8 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at F16, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek MoE 16B ChatDeepSeek · 16B (2.8B active)F1640.1 GB55.1 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at F16, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.LFM2 8B-A1BLiquid AI · 8.3B (1.5B active)F1623.1 GB85.0 tok/sRuns great
39Basic quality (D) · 39 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.InternLM2 Chat 20BShanghai AI Lab · 20BQ4_K_M20.1 GB43.6 tok/sRuns great
42Fair quality (C) · 42 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Mistral-NeMo-Minitron 8B InstructNVIDIA · 8.4BQ8_016.6 GB52.4 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at F16, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Unlimited OCR 2 (3B-A0.6B)Baidu · 3.3B (0.6B active)F1613.5 GB136 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Phi-4 Multimodal InstructMicrosoft · 5.6BQ8_013.3 GB79.1 tok/sRuns great
42Fair quality (C) · 42 of 100: how capable the model is at F16, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Hunyuan 4B InstructTencent · 4.2BF1616.0 GB48.1 tok/sRuns great
42Fair quality (C) · 42 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Llama 3 8B InstructMeta · 8.0BQ8_015.9 GB54.9 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Nous Hermes 2 SOLAR 10.7BNous Research · 11BQ8_018.5 GB41.1 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.SOLAR 10.7B InstructUpstage · 11BQ8_018.5 GB41.1 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at F16, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Jamba Reasoning 3BAI21 Labs · 3.2BF1612.8 GB63.5 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at F16, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.LFM2.5-VL 3BLiquid AI · 3.1BF1612.7 GB65.1 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Llama 2 70B ChatMeta · 69BQ4_K_M49.8 GB12.6 tok/sRuns well
42Fair quality (C) · 42 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Command R7B ArabicCohere · 8.0BQ8_015.2 GB54.9 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at F16, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.North Micro VisionCohere · 2.5BF1612.3 GB81.9 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek Prover V2 7BDeepSeek · 6.9BQ8_017.7 GB63.8 tok/sRuns great
44Fair quality (C) · 44 of 100: how capable the model is at F16, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.LFM2.5 2.6BLiquid AI · 2.7BF1611.9 GB75.2 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.InternLM2 Math Plus 20BShanghai AI Lab · 20BQ4_K_M20.1 GB43.6 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at F16, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Phi-3.5 Mini InstructMicrosoft · 3.8BF1617.2 GB53.2 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.CogVLM2 Llama 3 Chat 19BZhipu AI · 20BQ4_K_M22.5 GB44.4 tok/sRuns great
65Good quality (B) · 65 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Step 3.7 Flash (198B-A11B)StepFun · 201B (11.0B active)Q2_K87.3 GB78.3 tok/sSlowfar over GPU memory limit
38Basic quality (D) · 38 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.StarCoder2 15B InstructBigCode · 16BQ5_K_M18.4 GB44.8 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Yi Coder 9B Chat01.AI · 8.8BQ8_016.5 GB49.9 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at F16, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.AFM 4.5BArcee AI · 4.6BF1616.2 GB44.0 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Apertus 8B InstructSwiss AI · 8.1BQ8_016.0 GB54.7 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Aya Expanse 8BCohere · 8.0BQ8_015.9 GB54.9 tok/sRuns great
42Fair quality (C) · 42 of 100: how capable the model is at F16, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.SmolLM3 3BHugging Face · 3.1BF1613.1 GB65.9 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Yi 34B Chat01.AI · 34BQ4_K_M28.3 GB25.2 tok/sRuns well
33Basic quality (D) · 33 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Qwen1.5 14B ChatAlibaba · 14BQ6_K24.7 GB42.4 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Granite 3.0 8B InstructIBM · 8.2BQ8_015.7 GB53.9 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.InternLM2.5 7B ChatShanghai AI Lab · 7.7BQ8_015.6 GB56.9 tok/sRuns great
64Good quality (B) · 64 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Step 3.5 Flash (196B-A11B)StepFun · 199B (11.0B active)Q2_K86.5 GB78.3 tok/sSlowfar over GPU memory limit
42Fair quality (C) · 42 of 100: how capable the model is at F16, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Jamba2 3BAI21 Labs · 3.0BF1612.4 GB67.0 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Granite Code 34B InstructIBM · 34BQ4_K_M27.3 GB25.7 tok/sRuns well
38Basic quality (D) · 38 of 100: how capable the model is at F16, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Phi-3.5 Vision InstructMicrosoft · 4.2BF1617.9 GB48.9 tok/sRuns great
24Basic quality (D) · 24 of 100: how capable the model is at F16, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Qwen1.5 MoE A2.7B ChatAlibaba · 14B (2.7B active)F1636.6 GB56.6 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at F16, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Granite 4.0 Micro (3B)IBM · 3.4BF1613.8 GB59.7 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Code Llama 34B InstructMeta · 34BQ4_K_M28.6 GB25.7 tok/sRuns well
38Basic quality (D) · 38 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.LLM4Decompile 9B V2LLM4Binary · 8.8BQ8_016.5 GB49.9 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at F16, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Unlimited OCR (3B-A0.6B)Baidu · 3.3B (0.6B active)F1613.5 GB136 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Dolphin 2.9 Llama 3 8BCognitive Computations · 8.0BQ8_015.9 GB54.9 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.EXAONE 3.0 7.8B InstructLG AI Research · 7.8BQ8_015.2 GB56.4 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Mistral 7B Instruct v0.3Mistral AI · 7.3BQ8_015.1 GB60.8 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Zamba2 7B InstructZyphra · 7.4BQ8_014.9 GB59.9 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.MiniCPM3 4BOpenBMB · 4.1BF1614.8 GB49.9 tok/sRuns great
42Fair quality (C) · 42 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.MiniCPM-OCR 1.0 1.3BOpenBMB · 1.3BF169.1 GB156 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Gemma 3n E2BGoogle · 5.4BF1617.4 GB102 tok/sRuns great
37Basic quality (D) · 37 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.S5 nanoMaritaca AI · 4.7BF1615.9 GB43.6 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Phi-3 Mini 128K InstructMicrosoft · 3.8BF1617.2 GB53.2 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Falcon3 Mamba 7B InstructTII · 7.3BQ8_014.1 GB60.6 tok/sRuns great
42Fair quality (C) · 42 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.PaddleOCR-VL 1.6 0.9BBaidu · 1.0BF168.4 GB212 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Voxtral Mini 3BMistral AI · 4.7BF1616.7 GB43.4 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.5 2BAlibaba · 2.3BF1611.0 GB89.5 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.CodeGemma 7BGoogle · 8.5BQ8_019.2 GB51.6 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Granite Code 20B InstructIBM · 20BQ4_K_M18.8 GB43.1 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.OLMo 2 7B InstructAi2 · 7.3BQ8_016.2 GB60.4 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Yi 1.5 9B Chat01.AI · 8.8BQ8_016.1 GB49.9 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.TMax 2BAi2 · 1.9BF1610.2 GB108 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.TeleOCRXingChen-AGI · 1.4BF1610.1 GB143 tok/sRuns great
38Basic quality (D) · 38 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Granite 4.0 H MicroIBM · 3.2BF1612.8 GB63.7 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Phi-3 Mini 4K InstructMicrosoft · 3.8BF1615.6 GB53.2 tok/sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.MiniCPM-V 4.7 1BOpenBMB · 1.6BF169.6 GB131 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Hermes 2 Pro Mistral 7BNous Research · 7.2BQ8_015.1 GB60.9 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Starling LM 7B BetaNexusflow · 7.2BQ8_015.1 GB60.9 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at F16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Phi-3 Vision 128K InstructMicrosoft · 4.2BF1617.9 GB48.9 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek Coder 6.7B InstructDeepSeek · 6.7BQ8_017.8 GB65.4 tok/sRuns great
29Basic quality (D) · 29 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.Code Llama 13B InstructMeta · 13BQ6_K23.7 GB46.1 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5-Math 7B InstructAlibaba · 7.6BQ8_014.7 GB57.8 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.Codestral Mamba 7BMistral AI · 7.3BQ8_014.1 GB60.5 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.RecurrentGemma 9B ITGoogle · 9.6BQ8_016.7 GB45.8 tok/sRuns great
35Basic quality (D) · 35 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.OpenChat 3.5 (0106)OpenChat · 7.2BQ8_015.1 GB60.9 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.Aya 23 8BCohere · 8.0BQ8_015.9 GB54.9 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at F16, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.Falcon-H1 3B InstructTII · 3.2BF1612.9 GB64.5 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at F16, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 Coder 3B InstructAlibaba · 3.1BF1612.8 GB65.7 tok/sRuns great
35Basic quality (D) · 35 of 100: how capable the model is at F16, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.2 3B InstructMeta · 3.2BF1613.7 GB63.3 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.Mathstral 7BMistral AI · 7.3BQ8_015.1 GB60.8 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.Mistral 7B Instruct v0.2Mistral AI · 7.2BQ8_015.1 GB60.9 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.Nous Hermes 2 Mistral 7B DPONous Research · 7.2BQ8_015.1 GB60.9 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.OpenChat 3.5 (1210)OpenChat · 7.2BQ8_015.1 GB60.9 tok/sRuns great
35Basic quality (D) · 35 of 100: how capable the model is at F16, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.Raon 1 nanoKRAFTON · 3.5BF1613.6 GB57.5 tok/sRuns great
32Basic quality (D) · 32 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.LLM4Decompile 6.7B V2LLM4Binary · 6.7BQ8_017.8 GB65.4 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at F16, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.Qwen3-VL 2B InstructAlibaba · 2.1BF1611.6 GB95.4 tok/sRuns great
31Basic quality (D) · 31 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.StarCoder2 15B (base)BigCode · 16BQ5_K_M18.4 GB44.8 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at F16, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5-VL 3B InstructAlibaba · 3.8BF1614.2 GB54.2 tok/sRuns great
32Basic quality (D) · 32 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.Granite Code 8B Instruct (128K)IBM · 8.1BQ8_016.1 GB54.7 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at F16, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 3B InstructAlibaba · 3.1BF1612.8 GB65.7 tok/sRuns great
32Basic quality (D) · 32 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.InternLM2 Math Plus 7BShanghai AI Lab · 7.7BQ8_015.6 GB56.9 tok/sRuns great
32Basic quality (D) · 32 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.OpenHermes 2.5 Mistral 7BTeknium · 7.2BQ8_015.1 GB60.9 tok/sRuns great
32Basic quality (D) · 32 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.Starling LM 7B AlphaBerkeley · 7.2BQ8_015.1 GB60.9 tok/sRuns great
30Basic quality (D) · 30 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.SQLCoder 7B v2Defog · 6.7BQ8_017.8 GB65.4 tok/sRuns great
34Basic quality (D) · 34 of 100: how capable the model is at F16, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.LFM2 2.6BLiquid AI · 2.6BF1611.6 GB79.0 tok/sRuns great
27Basic quality (D) · 27 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.Orca 2 13BMicrosoft · 13BQ6_K20.4 GB46.1 tok/sRuns great
36Basic quality (D) · 36 of 100: how capable the model is at F16, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.MiniCPM5 1BOpenBMB · 1.1BF168.7 GB188 tok/sRuns great
32Basic quality (D) · 32 of 100: how capable the model is at F16, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.Hermes 3 Llama 3.2 3BNous Research · 3.2BF1613.7 GB63.3 tok/sRuns great
28Basic quality (D) · 28 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.Gemma 1.1 7B ITGoogle · 8.5BQ8_019.2 GB51.6 tok/sRuns great
25Basic quality (D) · 25 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.Baichuan 2 13B ChatBaichuan · 14BQ6_K21.1 GB43.2 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at F16, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.OLMoE 1B-7B InstructAi2 · 6.9B (1.3B active)F1620.7 GB92.8 tok/sRuns great
30Basic quality (D) · 30 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.Mistral 7B Instruct v0.1Mistral AI · 7.2BQ8_015.1 GB60.9 tok/sRuns great
25Basic quality (D) · 25 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.Llama 2 13B ChatMeta · 13BQ6_K20.4 GB46.1 tok/sRuns great
30Basic quality (D) · 30 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.Falcon Mamba 7B InstructTII · 7.3BQ8_014.1 GB60.6 tok/sRuns great
27Basic quality (D) · 27 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.StableLM 2 12B ChatStability AI · 12BQ6_K17.2 GB49.5 tok/sRuns great
30Basic quality (D) · 30 of 100: how capable the model is at F16, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.Falcon 3 3B InstructTII · 3.2BF1613.6 GB62.9 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.Gemma 7B ITGoogle · 8.5BQ8_019.2 GB51.6 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.Qwen1.5 7B ChatAlibaba · 7.7BQ8_018.8 GB57.1 tok/sRuns great
28Basic quality (D) · 28 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.LLaVA 1.6 Mistral 7BLLaVA · 7.6BQ8_015.5 GB58.2 tok/sRuns great
30Basic quality (D) · 30 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5-Omni 3BAlibaba · 5.5BQ8_012.5 GB79.6 tok/sRuns great
28Basic quality (D) · 28 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.Neural Chat 7B v3.3Intel · 7.2BQ8_015.1 GB60.9 tok/sRuns great
30Basic quality (D) · 30 of 100: how capable the model is at F16, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.Granite 3.2 2B InstructIBM · 2.5BF1612.1 GB80.3 tok/sRuns great
30Basic quality (D) · 30 of 100: how capable the model is at F16, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.Granite 3.3 2BIBM · 2.5BF1612.1 GB80.3 tok/sRuns great
30Basic quality (D) · 30 of 100: how capable the model is at F16, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.EXAONE 3.5 2.4BLG AI Research · 2.4BF1611.8 GB84.3 tok/sRuns great
32Basic quality (D) · 32 of 100: how capable the model is at F16, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.LFM2.5 1.2B InstructLiquid AI · 1.2BF168.8 GB174 tok/sRuns great
30Basic quality (D) · 30 of 100: how capable the model is at F16, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 1.7BAlibaba · 2.0BF1611.4 GB100 tok/sRuns great
25Basic quality (D) · 25 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.StarCoder 15BBigCode · 16BQ5_K_M17.8 GB45.2 tok/sRuns great
28Basic quality (D) · 28 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.Yi 1.5 6B Chat01.AI · 6.1BQ8_013.1 GB72.7 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.WizardMath 7B v1.1WizardLM · 7.2BQ8_015.1 GB60.9 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.Zephyr 7B BetaHugging Face · 7.2BQ8_015.1 GB60.9 tok/sRuns great
30Basic quality (D) · 30 of 100: how capable the model is at F16, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.EXAONE 4.0 1.2BLG AI Research · 1.3BF169.4 GB159 tok/sRuns great
24Basic quality (D) · 24 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.Code Llama 7B InstructMeta · 6.7BQ8_017.8 GB65.4 tok/sRuns great
28Basic quality (D) · 28 of 100: how capable the model is at F16, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.Granite 3.1 2B InstructIBM · 2.5BF1612.1 GB80.3 tok/sRuns great
28Basic quality (D) · 28 of 100: how capable the model is at F16, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.Gemma 2 2BGoogle · 2.6BF1612.1 GB77.8 tok/sRuns great
21Basic quality (D) · 21 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.LLaVA 1.5 13BLLaVA · 13BQ6_K20.7 GB45.0 tok/sRuns great
21Basic quality (D) · 21 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.WizardLM 13B v1.2WizardLM · 13BQ6_K20.4 GB46.1 tok/sRuns great
27Basic quality (D) · 27 of 100: how capable the model is at F16, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.WeVisOne VL 2BTencent · 2.4BF1612.2 GB83.2 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at F16, whatever it runs on. Match score 55/100 (quality, fit, speed and memory use) sets the ranking.Granite Vision 3.2 2BIBM · 3.0BF1613.0 GB68.2 tok/sRuns great
28Basic quality (D) · 28 of 100: how capable the model is at F16, whatever it runs on. Match score 55/100 (quality, fit, speed and memory use) sets the ranking.Granite 4.0 H 1BIBM · 1.5BF169.3 GB139 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at F16, whatever it runs on. Match score 55/100 (quality, fit, speed and memory use) sets the ranking.Granite 3.0 2B InstructIBM · 2.6BF1612.0 GB77.2 tok/sRuns great
24Basic quality (D) · 24 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 55/100 (quality, fit, speed and memory use) sets the ranking.StarCoder2 7BBigCode · 7.2BQ8_014.5 GB61.5 tok/sRuns great
19Basic quality (D) · 19 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 55/100 (quality, fit, speed and memory use) sets the ranking.Vicuna 13B v1.5LMSYS · 13BQ6_K20.4 GB46.1 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at F16, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.DeepScaleR 1.5B PreviewAgentica · 1.8BF1610.2 GB114 tok/sRuns great
22Basic quality (D) · 22 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.Orca 2 7BMicrosoft · 6.7BQ8_015.7 GB65.4 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at F16, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 Coder 1.5B InstructAlibaba · 1.5BF169.7 GB132 tok/sRuns great
24Basic quality (D) · 24 of 100: how capable the model is at F16, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.SmolVLM2 2.2B InstructHugging Face · 2.3BF1612.5 GB90.3 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at F16, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.Falcon-H1 1.5B InstructTII · 1.6BF169.7 GB131 tok/sRuns great
22Basic quality (D) · 22 of 100: how capable the model is at F16, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.Nemotron Mini 4B InstructNVIDIA · 4.2BF1615.3 GB48.5 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at F16, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.LFM2 1.2BLiquid AI · 1.2BF168.8 GB174 tok/sRuns great
24Basic quality (D) · 24 of 100: how capable the model is at F16, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.Qwen2-VL 2B InstructAlibaba · 2.2BF1611.0 GB91.9 tok/sRuns great
26Basic quality (D) · 26 of 100: how capable the model is at F16, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.5 0.8BAlibaba · 0.9BF168.2 GB233 tok/sRuns great
20Basic quality (D) · 20 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.Baichuan 2 7B ChatBaichuan · 7.5BQ8_016.5 GB58.7 tok/sRuns great
20Basic quality (D) · 20 of 100: how capable the model is at F16, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.Granite Code 3B Instruct (128K)IBM · 3.5BF1616.0 GB58.4 tok/sRuns great
20Basic quality (D) · 20 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.Llama 2 7B ChatMeta · 6.7BQ8_015.7 GB65.4 tok/sRuns great
24Basic quality (D) · 24 of 100: how capable the model is at F16, whatever it runs on. Match score 53/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 1.5B InstructAlibaba · 1.5BF169.7 GB132 tok/sRuns great
22Basic quality (D) · 22 of 100: how capable the model is at F16, whatever it runs on. Match score 53/100 (quality, fit, speed and memory use) sets the ranking.SmolVLM 2.2B InstructHugging Face · 2.3BF1612.5 GB90.3 tok/sRuns great
24Basic quality (D) · 24 of 100: how capable the model is at F16, whatever it runs on. Match score 53/100 (quality, fit, speed and memory use) sets the ranking.LFM2-VL 1.6BLiquid AI · 1.6BF169.6 GB129 tok/sRuns great
22Basic quality (D) · 22 of 100: how capable the model is at F16, whatever it runs on. Match score 53/100 (quality, fit, speed and memory use) sets the ranking.Zamba2 2.7B InstructZyphra · 2.7BF1612.1 GB76.4 tok/sRuns great
20Basic quality (D) · 20 of 100: how capable the model is at F16, whatever it runs on. Match score 53/100 (quality, fit, speed and memory use) sets the ranking.Stable Code Instruct 3BStability AI · 2.8BF1614.6 GB72.5 tok/sRuns great
20Basic quality (D) · 20 of 100: how capable the model is at F16, whatever it runs on. Match score 53/100 (quality, fit, speed and memory use) sets the ranking.Granite 3.1 3B-A800M InstructIBM · 3.3B (0.8B active)F1613.5 GB120 tok/sRuns great
22Basic quality (D) · 22 of 100: how capable the model is at F16, whatever it runs on. Match score 53/100 (quality, fit, speed and memory use) sets the ranking.Hunyuan 1.8B InstructTencent · 1.8BF1610.5 GB113 tok/sRuns great
20Basic quality (D) · 20 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 53/100 (quality, fit, speed and memory use) sets the ranking.Yi 6B Chat01.AI · 6.1BQ8_013.1 GB72.7 tok/sRuns great
22Basic quality (D) · 22 of 100: how capable the model is at F16, whatever it runs on. Match score 53/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek R1 Distill Qwen 1.5BDeepSeek · 1.8BF1610.2 GB114 tok/sRuns great
18Basic quality (D) · 18 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 53/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek LLM 7B ChatDeepSeek · 6.9BQ8_015.7 GB63.8 tok/sRuns great
22Basic quality (D) · 22 of 100: how capable the model is at F16, whatever it runs on. Match score 52/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.2 1B InstructMeta · 1.2BF169.1 GB164 tok/sRuns great
21Basic quality (D) · 21 of 100: how capable the model is at F16, whatever it runs on. Match score 52/100 (quality, fit, speed and memory use) sets the ranking.d1 2BLiquid AI · 1.9BF1610.3 GB106 tok/sRuns great
20Basic quality (D) · 20 of 100: how capable the model is at F16, whatever it runs on. Match score 52/100 (quality, fit, speed and memory use) sets the ranking.SmolLM2 1.7B InstructHugging Face · 1.7BF1611.4 GB119 tok/sRuns great
18Basic quality (D) · 18 of 100: how capable the model is at F16, whatever it runs on. Match score 52/100 (quality, fit, speed and memory use) sets the ranking.Granite 3.0 3B-A800M InstructIBM · 3.4B (0.8B active)F1613.4 GB120 tok/sRuns great
16Basic quality (D) · 16 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 52/100 (quality, fit, speed and memory use) sets the ranking.LLaVA 1.5 7BLLaVA · 7.1BQ8_016.0 GB62.4 tok/sRuns great
18Basic quality (D) · 18 of 100: how capable the model is at F16, whatever it runs on. Match score 52/100 (quality, fit, speed and memory use) sets the ranking.StarCoder2 3BBigCode · 3.0BF1612.7 GB67.0 tok/sRuns great
20Basic quality (D) · 20 of 100: how capable the model is at F16, whatever it runs on. Match score 52/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5-Math 1.5B InstructAlibaba · 1.5BF169.6 GB132 tok/sRuns great
20Basic quality (D) · 20 of 100: how capable the model is at F16, whatever it runs on. Match score 51/100 (quality, fit, speed and memory use) sets the ranking.Gemma 3 1BGoogle · 1.0BF168.4 GB203 tok/sRuns great
18Basic quality (D) · 18 of 100: how capable the model is at F16, whatever it runs on. Match score 51/100 (quality, fit, speed and memory use) sets the ranking.InternLM2.5 1.8B ChatShanghai AI Lab · 1.9BF1610.9 GB107 tok/sRuns great
16Basic quality (D) · 16 of 100: how capable the model is at F16, whatever it runs on. Match score 51/100 (quality, fit, speed and memory use) sets the ranking.StableLM Zephyr 3BStability AI · 2.8BF1613.3 GB72.5 tok/sRuns great
11Basic quality (D) · 11 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 51/100 (quality, fit, speed and memory use) sets the ranking.Falcon 40B InstructTII · 42BQ2_K23.3 GB34.0 tok/sRuns great
14Basic quality (D) · 14 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 51/100 (quality, fit, speed and memory use) sets the ranking.Vicuna 7B v1.5LMSYS · 6.7BQ8_015.7 GB65.4 tok/sRuns great
16Basic quality (D) · 16 of 100: how capable the model is at F16, whatever it runs on. Match score 51/100 (quality, fit, speed and memory use) sets the ranking.Phi-2 2.7BMicrosoft · 2.8BF1612.6 GB73.1 tok/sRuns great
14Basic quality (D) · 14 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 51/100 (quality, fit, speed and memory use) sets the ranking.OLMo 7B InstructAi2 · 6.9BQ8_014.7 GB64.0 tok/sRuns great
17Basic quality (D) · 17 of 100: how capable the model is at F16, whatever it runs on. Match score 50/100 (quality, fit, speed and memory use) sets the ranking.GLiDE 2BFastino · 1.9BF1610.2 GB108 tok/sRuns great
16Basic quality (D) · 16 of 100: how capable the model is at F16, whatever it runs on. Match score 50/100 (quality, fit, speed and memory use) sets the ranking.Yi Coder 1.5B Chat01.AI · 1.5BF1610.9 GB137 tok/sRuns great
18Basic quality (D) · 18 of 100: how capable the model is at F16, whatever it runs on. Match score 50/100 (quality, fit, speed and memory use) sets the ranking.LFM2 700MLiquid AI · 0.7BF167.9 GB274 tok/sRuns great
16Basic quality (D) · 16 of 100: how capable the model is at F16, whatever it runs on. Match score 50/100 (quality, fit, speed and memory use) sets the ranking.Falcon 3 1B InstructTII · 1.7BF1610.3 GB122 tok/sRuns great
16Basic quality (D) · 16 of 100: how capable the model is at F16, whatever it runs on. Match score 50/100 (quality, fit, speed and memory use) sets the ranking.Spark X2.6 1.7BSparkLLM · 1.7BF169.9 GB119 tok/sRuns great
16Basic quality (D) · 16 of 100: how capable the model is at F16, whatever it runs on. Match score 50/100 (quality, fit, speed and memory use) sets the ranking.Qwen2 1.5B InstructAlibaba · 1.5BF169.7 GB132 tok/sRuns great
12Basic quality (D) · 12 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 50/100 (quality, fit, speed and memory use) sets the ranking.InternLM 7BShanghai AI Lab · 7.3BQ8_015.2 GB60.2 tok/sRuns great
14Basic quality (D) · 14 of 100: how capable the model is at F16, whatever it runs on. Match score 49/100 (quality, fit, speed and memory use) sets the ranking.CodeGemma 2BGoogle · 2.5BF1611.5 GB80.9 tok/sRuns great
14Basic quality (D) · 14 of 100: how capable the model is at F16, whatever it runs on. Match score 49/100 (quality, fit, speed and memory use) sets the ranking.SmolLM 1.7B InstructHugging Face · 1.7BF1610.2 GB119 tok/sRuns great
14Basic quality (D) · 14 of 100: how capable the model is at F16, whatever it runs on. Match score 49/100 (quality, fit, speed and memory use) sets the ranking.OLMo 2 1B InstructAi2 · 1.5BF169.9 GB137 tok/sRuns great
15Basic quality (D) · 15 of 100: how capable the model is at F16, whatever it runs on. Match score 49/100 (quality, fit, speed and memory use) sets the ranking.surya ocr 2.5Datalab · 0.7BF167.8 GB294 tok/sRuns great
14Basic quality (D) · 14 of 100: how capable the model is at F16, whatever it runs on. Match score 49/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 0.6BAlibaba · 0.8BF168.8 GB271 tok/sRuns great
12Basic quality (D) · 12 of 100: how capable the model is at F16, whatever it runs on. Match score 49/100 (quality, fit, speed and memory use) sets the ranking.Gemma 2B ITGoogle · 2.5BF1611.5 GB80.9 tok/sRuns great
12Basic quality (D) · 12 of 100: how capable the model is at F16, whatever it runs on. Match score 48/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek Coder 1.3B InstructDeepSeek · 1.4BF1610.7 GB150 tok/sRuns great
12Basic quality (D) · 12 of 100: how capable the model is at F16, whatever it runs on. Match score 48/100 (quality, fit, speed and memory use) sets the ranking.LLM4Decompile 1.3B V2LLM4Binary · 1.4BF1610.7 GB150 tok/sRuns great
14Basic quality (D) · 14 of 100: how capable the model is at F16, whatever it runs on. Match score 48/100 (quality, fit, speed and memory use) sets the ranking.Falcon-H1 0.5B InstructTII · 0.5BF167.6 GB391 tok/sRuns great
14Basic quality (D) · 14 of 100: how capable the model is at F16, whatever it runs on. Match score 48/100 (quality, fit, speed and memory use) sets the ranking.LFM2.5 350MLiquid AI · 0.4BF167.2 GB580 tok/sRuns great
10Basic quality (D) · 10 of 100: how capable the model is at F16, whatever it runs on. Match score 47/100 (quality, fit, speed and memory use) sets the ranking.StableLM 2 1.6B ChatStability AI · 1.6BF1610.4 GB124 tok/sRuns great
10Basic quality (D) · 10 of 100: how capable the model is at F16, whatever it runs on. Match score 47/100 (quality, fit, speed and memory use) sets the ranking.StableLM 2 Zephyr 1.6BStability AI · 1.6BF1610.4 GB124 tok/sRuns great
12Basic quality (D) · 12 of 100: how capable the model is at F16, whatever it runs on. Match score 47/100 (quality, fit, speed and memory use) sets the ranking.LFM2-VL 450MLiquid AI · 0.5BF167.4 GB451 tok/sRuns great
12Basic quality (D) · 12 of 100: how capable the model is at F16, whatever it runs on. Match score 47/100 (quality, fit, speed and memory use) sets the ranking.LFM2 350MLiquid AI · 0.4BF167.2 GB580 tok/sRuns great
12Basic quality (D) · 12 of 100: how capable the model is at F16, whatever it runs on. Match score 47/100 (quality, fit, speed and memory use) sets the ranking.Granite 4.0 H 350MIBM · 0.3BF167.1 GB597 tok/sRuns great
11Basic quality (D) · 11 of 100: how capable the model is at F16, whatever it runs on. Match score 47/100 (quality, fit, speed and memory use) sets the ranking.GLiDE 0.8BFastino · 0.8BF168.0 GB271 tok/sRuns great
6Basic quality (D) · 6 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 47/100 (quality, fit, speed and memory use) sets the ranking.Falcon 7B InstructTII · 7.2BQ8_014.0 GB61.0 tok/sRuns great
10Basic quality (D) · 10 of 100: how capable the model is at F16, whatever it runs on. Match score 46/100 (quality, fit, speed and memory use) sets the ranking.SmolVLM 500M InstructHugging Face · 0.5BF167.7 GB398 tok/sRuns great
10Basic quality (D) · 10 of 100: how capable the model is at F16, whatever it runs on. Match score 46/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 0.5B InstructAlibaba · 0.5BF167.4 GB415 tok/sRuns great
10Basic quality (D) · 10 of 100: how capable the model is at F16, whatever it runs on. Match score 46/100 (quality, fit, speed and memory use) sets the ranking.LFM2.5 230MLiquid AI · 0.2BF166.9 GB883 tok/sRuns great
8Basic quality (D) · 8 of 100: how capable the model is at F16, whatever it runs on. Match score 46/100 (quality, fit, speed and memory use) sets the ranking.Hunyuan 0.5B InstructTencent · 0.5BF168.2 GB376 tok/sRuns great
8Basic quality (D) · 8 of 100: how capable the model is at F16, whatever it runs on. Match score 45/100 (quality, fit, speed and memory use) sets the ranking.Qwen2.5 Coder 0.5B InstructAlibaba · 0.5BF167.4 GB415 tok/sRuns great
8Basic quality (D) · 8 of 100: how capable the model is at F16, whatever it runs on. Match score 45/100 (quality, fit, speed and memory use) sets the ranking.SmolLM2 360M InstructHugging Face · 0.4BF167.4 GB564 tok/sRuns great
6Basic quality (D) · 6 of 100: how capable the model is at F16, whatever it runs on. Match score 45/100 (quality, fit, speed and memory use) sets the ranking.TinyLlama 1.1B ChatTinyLlama · 1.1BF168.6 GB185 tok/sRuns great
6Basic quality (D) · 6 of 100: how capable the model is at F16, whatever it runs on. Match score 45/100 (quality, fit, speed and memory use) sets the ranking.Qwen2 0.5B InstructAlibaba · 0.5BF167.4 GB415 tok/sRuns great
6Basic quality (D) · 6 of 100: how capable the model is at F16, whatever it runs on. Match score 45/100 (quality, fit, speed and memory use) sets the ranking.ERNIE 4.5 0.3BBaidu · 0.4BF167.2 GB564 tok/sRuns great
6Basic quality (D) · 6 of 100: how capable the model is at F16, whatever it runs on. Match score 44/100 (quality, fit, speed and memory use) sets the ranking.SmolVLM 256M InstructHugging Face · 0.3BF167.1 GB781 tok/sRuns great
6Basic quality (D) · 6 of 100: how capable the model is at F16, whatever it runs on. Match score 44/100 (quality, fit, speed and memory use) sets the ranking.Gemma 3 270MGoogle · 0.3BF166.9 GB752 tok/sRuns great
5Basic quality (D) · 5 of 100: how capable the model is at F16, whatever it runs on. Match score 44/100 (quality, fit, speed and memory use) sets the ranking.SmolLM 360M InstructHugging Face · 0.4BF167.2 GB564 tok/sRuns great
4Basic quality (D) · 4 of 100: how capable the model is at F16, whatever it runs on. Match score 43/100 (quality, fit, speed and memory use) sets the ranking.SmolLM2 135M InstructHugging Face · 0.1BF166.8 GB1,562 tok/sRuns great
3Basic quality (D) · 3 of 100: how capable the model is at F16, whatever it runs on. Match score 43/100 (quality, fit, speed and memory use) sets the ranking.Falcon-H1 Tiny 90M InstructTII · 0.1BF166.6 GB2,257 tok/sRuns great
2Basic quality (D) · 2 of 100: how capable the model is at F16, whatever it runs on. Match score 43/100 (quality, fit, speed and memory use) sets the ranking.SmolLM 135M InstructHugging Face · 0.1BF166.7 GB1,562 tok/sRuns great
23Basic quality (D) · 23 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 41/100 (quality, fit, speed and memory use) sets the ranking.Falcon 180B ChatTII · 180BQ2_K78.5 GB6.7 tok/sSlowover GPU memory limit
69Good quality (B) · 69 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.Kimi K3.1 Pro (2.8T)Moonshot AI · 2,780B (104.0B active)Q2_K1,119 GB—Won't run
68Good quality (B) · 68 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V4.1 Flash 1017 (552B-A16B)DeepSeek · 763B (16.0B active)Q2_K312 GB—Won't run
68Good quality (B) · 68 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.GLM-5.3 (753B-A40B)Zhipu AI · 753B (40.0B active)Q2_K308 GB—Won't run
68Good quality (B) · 68 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.8 Max (2.4T-A95B)Alibaba · 2,446B (95.0B active)Q2_K986 GB—Won't run
68Good quality (B) · 68 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.Kimi K3 (2.8T)Moonshot AI · 2,780B (104.0B active)Q2_K1,119 GB—Won't run
68Good quality (B) · 68 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.LongCat 2.5 (1.6T-A48B)Meituan · 1,780B (48.0B active)Q2_K719 GB—Won't run
68Good quality (B) · 68 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.MiMo V2.6 Pro (1T-A42B)Xiaomi · 1,024B (42.0B active)Q2_K417 GB—Won't run
68Good quality (B) · 68 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.Ling 3.1 Flash 1019 (560B-A25B)Ant Group (inclusionAI) · 572B (25.0B active)Q2_K235 GB—Won't run
67Good quality (B) · 67 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.MiniMax M3.1 Flash (428B-A23B)MiniMax · 429B (23.0B active)Q2_K179 GB—Won't run
67Good quality (B) · 67 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.Nex N2.6 (321B-A18B)Nex-AGI · 321B (18.0B active)Q2_K135 GB—Won't run
67Good quality (B) · 67 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.Step 5 Preview (600B-A27B)StepFun · 604B (27.0B active)Q2_K249 GB—Won't run
67Good quality (B) · 67 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.GLM-5.3 Flash (320B-A18B)Zhipu AI · 321B (18.0B active)Q2_K135 GB—Won't run
67Good quality (B) · 67 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.MiMo V2.6 Flash (310B-A15B)Xiaomi · 311B (15.0B active)Q2_K131 GB—Won't run
67Good quality (B) · 67 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V4.1 Flash (552B-A16B)DeepSeek · 763B (16.0B active)Q2_K312 GB—Won't run
67Good quality (B) · 67 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.Mistral Large 4 (1T-A52B)Mistral AI · 1,053B (52.0B active)Q2_K428 GB—Won't run
67Good quality (B) · 67 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.GLM-5.2 (753B-A40B)Zhipu AI · 753B (40.0B active)Q2_K308 GB—Won't run
67Good quality (B) · 67 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.Hunyuan Hy4 Preview (780B)Tencent · 780B (55.0B active)Q2_K319 GB—Won't run
67Good quality (B) · 67 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.Muse Spark Open (816B-A42B)Meta · 822B (42.0B active)Q2_K336 GB—Won't run
67Good quality (B) · 67 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V4 Pro (1.6T-A49B)DeepSeek · 1,651B (49.0B active)Q2_K667 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.Kimi K2.6 (1T)Moonshot AI · 1,027B (32.0B active)Q2_K418 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.Ling 3.1 Flash (560B-A25B)Ant Group (inclusionAI) · 572B (25.0B active)Q2_K235 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.MiMo V2.5 Pro (1T-A42B)Xiaomi · 1,023B (42.0B active)Q2_K416 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.Intern-S2 (397B-A17B)Shanghai AI Lab · 403B (17.0B active)Q2_K168 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.GLM-5.1 (744B-A40B)Zhipu AI · 754B (40.0B active)Q2_K309 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Nex N2 Pro (397B-A17B)Nex-AGI · 397B (17.0B active)Q2_K165 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V4 Pro (April 2026)DeepSeek · 1,599B (49.0B active)Q2_K646 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Naive N0.5 Flash (309B-A15.5B)Naive AI · 309B (15.5B active)Q2_K130 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Qwen3.5 397B-A17BAlibaba · 403B (17.0B active)Q2_K168 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.MiniMax M3 (427B)MiniMax · 427B (25.0B active)Q2_K178 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.GLM-5 (744B-A40B)Zhipu AI · 754B (40.0B active)Q2_K309 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.K2-Horizon (375B-A23B)Institute of Foundation Models · 379B (23.0B active)Q2_K160 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Hunyuan Hy3 (295B-A21B)Tencent · 299B (21.0B active)Q2_K129 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Beam 1.0 (501B-A23B)Reflection AI · 501B (23.4B active)Q2_K207 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Kimi K2.5 (1T)Moonshot AI · 1,027B (32.0B active)Q2_K418 GB—Won't run
66Good quality (B) · 66 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Kimi K2.7 Code (1T)Moonshot AI · 1,027B (32.0B active)Q2_K418 GB—Won't run
65Good quality (B) · 65 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V4 Flash (284B-A13B)DeepSeek · 304B (13.0B active)Q2_K128 GB—Won't run
65Good quality (B) · 65 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.MiMo V2.5 (310B-A15B)Xiaomi · 311B (15.0B active)Q2_K131 GB—Won't run
65Good quality (B) · 65 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Ornith 1.5 397B-A17BDeepReinforce · 403B (17.0B active)Q2_K168 GB—Won't run
65Good quality (B) · 65 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V3.2 (671B)DeepSeek · 685B (37.0B active)Q2_K281 GB—Won't run
65Good quality (B) · 65 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Inkling Small (276B-A12B)Thinking Machines Lab · 266B (12.0B active)Q2_K113 GB—Won't run
65Good quality (B) · 65 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.GLM-4.7 (355B-A32B)Zhipu AI · 358B (32.0B active)Q2_K153 GB—Won't run
65Good quality (B) · 65 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Inkling (975B-A41B)Thinking Machines Lab · 952B (41.0B active)Q2_K388 GB—Won't run
65Good quality (B) · 65 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.MiniMax M2.7 (230B-A10B)MiniMax · 229B (10.0B active)Q2_K99.9 GB—Won't run
65Good quality (B) · 65 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Nemotron 3 Ultra 550B-A55BNVIDIA · 561B (55.0B active)Q2_K231 GB—Won't run
65Good quality (B) · 65 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek R1 0528 (671B)DeepSeek · 685B (37.0B active)Q2_K281 GB—Won't run
65Good quality (B) · 65 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Kimi K2 Thinking (1T)Moonshot AI · 1,026B (32.0B active)Q2_K417 GB—Won't run
64Good quality (B) · 64 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V3.2 ExpDeepSeek · 685B (37.0B active)Q2_K281 GB—Won't run
64Good quality (B) · 64 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Ling 2.6 1TAnt Group (inclusionAI) · 1,025B (63.0B active)Q2_K417 GB—Won't run
64Good quality (B) · 64 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.MiniMax M2.5 (230B-A10B)MiniMax · 229B (10.0B active)Q2_K99.9 GB—Won't run
64Good quality (B) · 64 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek R1 (671B)DeepSeek · 671B (37.0B active)Q2_K275 GB—Won't run
64Good quality (B) · 64 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V4 Flash (April 2026)DeepSeek · 291B (13.0B active)Q2_K123 GB—Won't run
64Good quality (B) · 64 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V3.1 TerminusDeepSeek · 685B (37.0B active)Q2_K281 GB—Won't run
63Good quality (B) · 63 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V3.1 (671B)DeepSeek · 685B (37.0B active)Q2_K281 GB—Won't run
63Good quality (B) · 63 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Kimi K2 Instruct 0905 (1T)Moonshot AI · 1,026B (32.0B active)Q2_K418 GB—Won't run
63Good quality (B) · 63 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.MiniMax M2.1 (230B-A10B)MiniMax · 229B (10.0B active)Q2_K99.9 GB—Won't run
63Good quality (B) · 63 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 235B-A22B Thinking 2507Alibaba · 235B (22.0B active)Q2_K102 GB—Won't run
63Good quality (B) · 63 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 Coder 480B-A35BAlibaba · 480B (35.0B active)Q2_K201 GB—Won't run
63Good quality (B) · 63 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.MiMo V2 Flash (309B-A15B)Xiaomi · 310B (15.0B active)Q2_K131 GB—Won't run
63Good quality (B) · 63 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Ling 1TAnt Group (inclusionAI) · 1,000B (50.0B active)Q2_K409 GB—Won't run
63Good quality (B) · 63 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Mistral Large 3 (675B)Mistral AI · 675B (41.0B active)Q2_K277 GB—Won't run
63Good quality (B) · 63 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 235B-A22B Instruct 2507Alibaba · 235B (22.0B active)Q2_K102 GB—Won't run
63Good quality (B) · 63 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Solar Open 2 250BUpstage · 250B (15.0B active)Q2_K108 GB—Won't run
62Good quality (B) · 62 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Kimi K2 Instruct (July 2025)Moonshot AI · 1,026B (32.0B active)Q2_K417 GB—Won't run
62Good quality (B) · 62 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.MiniMax M2 (230B-A10B)MiniMax · 229B (10.0B active)Q2_K99.9 GB—Won't run
62Good quality (B) · 62 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Qwen3-VL 235B-A22B InstructAlibaba · 236B (22.0B active)Q2_K102 GB—Won't run
61Good quality (B) · 61 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Command A+ (218B-A25B)Cohere · 219B (25.0B active)Q2_K94.1 GB—Won't run
61Good quality (B) · 61 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.K-EXAONE 236B-A23BLG AI Research · 237B (23.0B active)Q2_K102 GB—Won't run
61Good quality (B) · 61 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.LongCat Flash Chat (560B)Meituan · 562B (27.0B active)Q2_K231 GB—Won't run
61Good quality (B) · 61 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 235B-A22BAlibaba · 235B (22.0B active)Q2_K102 GB—Won't run
60Good quality (B) · 60 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.InternVL3.5 241B-A28BShanghai AI Lab · 241B (28.0B active)Q2_K104 GB—Won't run
60Good quality (B) · 60 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.1 Nemotron Ultra 253B v1NVIDIA · 253BQ2_K112 GB—Won't run
58Good quality (B) · 58 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.Baichuan M3 235B-A22BBaichuan · 235B (22.0B active)Q2_K102 GB—Won't run
58Good quality (B) · 58 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.GLM-4.6 (355B-A32B)Zhipu AI · 357B (32.0B active)Q2_K152 GB—Won't run
58Good quality (B) · 58 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V3 0324 (671B)DeepSeek · 685B (37.0B active)Q2_K281 GB—Won't run
58Good quality (B) · 58 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.MiniMax-M1 80KMiniMax · 456B (45.9B active)Q2_K189 GB—Won't run
57Good quality (B) · 57 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.GLM-4.5 (355B-A32B)Zhipu AI · 358B (32.0B active)Q2_K153 GB—Won't run
57Good quality (B) · 57 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.MiniMax-M1 40KMiniMax · 456B (45.9B active)Q2_K189 GB—Won't run
56Good quality (B) · 56 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.Trinity Large Thinking (400B-A13B)Arcee AI · 399B (13.0B active)Q2_K167 GB—Won't run
56Good quality (B) · 56 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.ERNIE 4.5 300B-A47BBaidu · 300B (47.0B active)Q2_K128 GB—Won't run
56Good quality (B) · 56 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.Hermes 4 405BNous Research · 406BQ2_K173 GB—Won't run
54Fair quality (C) · 54 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.NorthCUA Probeipo-ai · 397B (17.0B active)Q2_K165 GB—Won't run
53Fair quality (C) · 53 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.MiniMax-Text-01MiniMax · 456B (45.9B active)Q2_K189 GB—Won't run
53Fair quality (C) · 53 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.Tülu 3 405BAi2 · 406BQ2_K173 GB—Won't run
52Fair quality (C) · 52 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V3 (671B)DeepSeek · 671B (37.0B active)Q2_K275 GB—Won't run
51Fair quality (C) · 51 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 6/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek Coder V2 (236B)DeepSeek · 236B (21.0B active)Q2_K101 GB—Won't run
50Fair quality (C) · 50 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 6/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek Coder V2 Instruct (0614)DeepSeek · 236B (21.0B active)Q2_K101 GB—Won't run
50Fair quality (C) · 50 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 6/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V2.5 1210DeepSeek · 236B (21.0B active)Q2_K101 GB—Won't run
50Fair quality (C) · 50 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 6/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.1 405B InstructMeta · 406BQ2_K173 GB—Won't run
50Fair quality (C) · 50 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 6/100 (quality, fit, speed and memory use) sets the ranking.Llama 4 Maverick (17B x 128E)Meta · 402B (17.0B active)Q2_K169 GB—Won't run
49Fair quality (C) · 49 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 6/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V2.5DeepSeek · 236B (21.0B active)Q2_K101 GB—Won't run
49Fair quality (C) · 49 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 6/100 (quality, fit, speed and memory use) sets the ranking.Jamba 1.5 LargeAI21 Labs · 399B (94.0B active)Q2_K166 GB—Won't run
43Fair quality (C) · 43 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 5/100 (quality, fit, speed and memory use) sets the ranking.DeepSeek V2 ChatDeepSeek · 236B (21.0B active)Q2_K101 GB—Won't run
35Basic quality (D) · 35 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 4/100 (quality, fit, speed and memory use) sets the ranking.Grok-1 (314B)xAI · 314B (86.0B active)Q2_K134 GB—Won't run
35Basic quality (D) · 35 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 4/100 (quality, fit, speed and memory use) sets the ranking.Snowflake Arctic InstructSnowflake · 479B (17.0B active)Q2_K198 GB—Won't run

Image models on the Apple M3 Ultra (96 GB)

Quality Model Quant Memory Speed Runs
92Excellent quality (A+) · 92 of 100: how capable the model is at FP16, whatever it runs on. Match score 89/100 (quality, fit, speed and memory use) sets the ranking.Krea 2 TurboKrea · 12.8BFP1637.4 GB18 sRuns well
84Great quality (A) · 84 of 100: how capable the model is at FP16, whatever it runs on. Match score 89/100 (quality, fit, speed and memory use) sets the ranking.FLUX.2 [klein] 9BBlack Forest Labs · 9.1BFP1637.1 GB7.6 sRuns great
89Excellent quality (A+) · 89 of 100: how capable the model is at FP16, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.Qwen-Image 2.1 TurboAlibaba · 7.1BFP1646.7 GB49 sSlow
92Excellent quality (A+) · 92 of 100: how capable the model is at FP16, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.Ideogram 4Ideogram · 18.6BFP1657.3 GB1.8 minSlow
93Excellent quality (A+) · 93 of 100: how capable the model is at FP16, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking.Qwen-Image 2.1Alibaba · 7.1BFP1657.0 GB4.2 minVery slow
84Great quality (A) · 84 of 100: how capable the model is at FP16, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking.Z-Image TurboAlibaba (Tongyi) · 6.2BFP1622.7 GB11 sRuns great
90Excellent quality (A+) · 90 of 100: how capable the model is at FP8, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking.FLUX.2 [dev]Black Forest Labs · 32.2BFP859.7 GB2.3 minVery slow
85Excellent quality (A+) · 85 of 100: how capable the model is at FP16, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking.ERNIE-Image TurboBaidu · 8.0BFP1626.4 GB14 sRuns well
70Great quality (A) · 70 of 100: how capable the model is at FP16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.FLUX.1 [schnell]Black Forest Labs · 11.9BFP1636.4 GB8.5 sRuns great
91Excellent quality (A+) · 91 of 100: how capable the model is at FP16, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.FLUX 3 Image KleinBlack Forest Labs · 5.6BFP1622.4 GB24 sRuns well
90Excellent quality (A+) · 90 of 100: how capable the model is at FP8, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.Cosmos3-Super Text2Image (64B)NVIDIA · 64.0BFP869.1 GB2.5 minVery slowtight fit
84Great quality (A) · 84 of 100: how capable the model is at FP8, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.Ming-Image 0.1 DesignAnt Group (inclusionAI) · 6.2BFP839.2 GB62 sSlow
91Excellent quality (A+) · 91 of 100: how capable the model is at FP8, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.MiniMax H3 ImageMiniMax · 33.1BFP838.6 GB40 sRuns wellencoder unloaded after use
78Great quality (A) · 78 of 100: how capable the model is at FP16, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.FLUX.2 [klein] 4BBlack Forest Labs · 3.9BFP1618.1 GB4.2 sRuns great
93Excellent quality (A+) · 93 of 100: how capable the model is at FP16, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.Fibo 2.0 DevBria AI · 8.1BFP1627.7 GB1.6 minSlow
88Excellent quality (A+) · 88 of 100: how capable the model is at FP16, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.ERNIE-ImageBaidu · 8.0BFP1634.0 GB1.6 minSlow
82Great quality (A) · 82 of 100: how capable the model is at FP16, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.FLUX.1 Krea [dev]Black Forest Labs · 11.9BFP1636.4 GB51 sSlow
80Great quality (A) · 80 of 100: how capable the model is at FP16, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.FLUX.1 [dev]Black Forest Labs · 11.9BFP1636.4 GB51 sSlow
89Excellent quality (A+) · 89 of 100: how capable the model is at Q8 GGUF, whatever it runs on. Match score 79/100 (quality, fit, speed and memory use) sets the ranking.Qwen-Image 2512Alibaba · 20.4BQ8 GGUF36.5 GB7.3 minVery slow
82Great quality (A) · 82 of 100: how capable the model is at FP16, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.FLUX.1 Kontext [dev]Black Forest Labs · 11.9BFP1637.0 GB1.6 minSlow
86Excellent quality (A+) · 86 of 100: how capable the model is at FP8, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Qwen-ImageAlibaba · 20.4BFP835.1 GB5.7 minVery slow
62Good quality (B) · 62 of 100: how capable the model is at FP16, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Stable Diffusion 3.5 Large TurboStability AI · 8.1BFP1630.0 GB6.8 sRuns great
82Great quality (A) · 82 of 100: how capable the model is at FP16, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.MOSS-Image SFTOpenMOSS · 5.1BFP1629.3 GB59 sSlow
76Great quality (A) · 76 of 100: how capable the model is at FP16, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.FLUX.1 Fill [dev]Black Forest Labs · 11.9BFP1636.4 GB53 sSlow
86Excellent quality (A+) · 86 of 100: how capable the model is at Q8 GGUF, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.FireRed-Image-Edit 1.1rednote (FireRed Team) · 20.4BQ8 GGUF33.9 GB6.5 minVery slow
74Great quality (A) · 74 of 100: how capable the model is at FP8, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.HiDream-I1 FastHiDream.ai · 17.1BFP833.7 GB38 sRuns well
78Great quality (A) · 78 of 100: how capable the model is at FP8, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.HiDream-I1 DevHiDream.ai · 17.1BFP833.7 GB65 sSlow
85Excellent quality (A+) · 85 of 100: how capable the model is at Q8 GGUF, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Qwen-Image-Edit 2511Alibaba · 20.4BQ8 GGUF33.9 GB6.5 minVery slow
82Great quality (A) · 82 of 100: how capable the model is at FP8, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.HiDream-I1 FullHiDream.ai · 17.1BFP833.7 GB2.9 minVery slow
78Great quality (A) · 78 of 100: how capable the model is at FP16, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Chroma1-HDLodestone Rock · 8.9BFP1630.0 GB85 sSlow
82Great quality (A) · 82 of 100: how capable the model is at FP16, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Z-ImageAlibaba (Tongyi) · 6.2BFP1622.7 GB72 sSlow
70Great quality (A) · 70 of 100: how capable the model is at FP16, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.CogView4 6BZhipu AI · 6.4BFP1633.9 GB76 sSlow
66Good quality (B) · 66 of 100: how capable the model is at FP16, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.Stable Diffusion 3.5 LargeStability AI · 8.1BFP1630.0 GB47 sSlow
58Good quality (B) · 58 of 100: how capable the model is at FP16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.DreamShaper XL v2 TurboLykon · 2.6BFP168.9 GB5.5 sRuns great
72Great quality (A) · 72 of 100: how capable the model is at FP16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.Iris-3BSperidlabs · 3.0BFP1617.3 GB68 sSlow
54Fair quality (C) · 54 of 100: how capable the model is at FP16, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.SDXL-Lightning (4-step)ByteDance · 2.6BFP168.9 GB3.0 sRuns great
58Good quality (B) · 58 of 100: how capable the model is at FP8, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.KolorsKuaishou · 2.6BFP811.0 GB15 sRuns well
50Fair quality (C) · 50 of 100: how capable the model is at FP16, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.PixArt-Σ XL 1024PixArt · 0.6BFP1612.8 GB6.3 sRuns great
58Good quality (B) · 58 of 100: how capable the model is at FP16, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.Stable Diffusion 3.5 MediumStability AI · 2.5BFP1618.4 GB25 sRuns well
66Good quality (B) · 66 of 100: how capable the model is at FP16, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.OmniGen2BAAI (VectorSpaceLab) · 4.0BFP1618.0 GB68 sSlow
62Good quality (B) · 62 of 100: how capable the model is at FP16, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Lumina-Image 2.0Shanghai AI Lab · 2.6BFP1612.7 GB30 sRuns well
60Good quality (B) · 60 of 100: how capable the model is at FP8, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Juggernaut XL v9RunDiffusion · 2.6BFP85.6 GB17 sRuns well
60Good quality (B) · 60 of 100: how capable the model is at FP8, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.RealVisXL V5.0SG161222 · 2.6BFP85.6 GB17 sRuns well
58Good quality (B) · 58 of 100: how capable the model is at FP8, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Animagine XL 4.0Cagliostro Lab · 2.6BFP85.6 GB16 sRuns well
58Good quality (B) · 58 of 100: how capable the model is at FP8, whatever it runs on. Match score 65/100 (quality, fit, speed and memory use) sets the ranking.Illustrious XL v0.1OnomaAI · 2.6BFP85.6 GB16 sRuns well
88Excellent quality (A+) · 88 of 100: how capable the model is at FP8, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.HunyuanImage 3.0 (80B-A13B)Tencent · 83.0BFP832.8 GB5.0 minVery slowCPU offload
50Fair quality (C) · 50 of 100: how capable the model is at FP8, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Stable Diffusion 3 MediumStability AI · 2.0BFP89.9 GB12 sRuns great
56Good quality (B) · 56 of 100: how capable the model is at FP8, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.Pony Diffusion V6 XLPurpleSmartAI · 2.6BFP85.6 GB14 sRuns well
54Fair quality (C) · 54 of 100: how capable the model is at FP8, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Animagine XL 3.1Cagliostro Lab · 2.6BFP85.6 GB16 sRuns well
60Good quality (B) · 60 of 100: how capable the model is at FP16, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Playground v2.5Playground AI · 2.6BFP168.9 GB32 sRuns well
58Good quality (B) · 58 of 100: how capable the model is at FP16, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.AuraFlow v0.3fal · 6.9BFP1618.5 GB59 sSlow
48Fair quality (C) · 48 of 100: how capable the model is at FP16, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Kandinsky 3Sber AI · 3.0BFP1625.9 GB38 sRuns well
53Fair quality (C) · 53 of 100: how capable the model is at FP8, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Stable Diffusion XL 1.0Stability AI · 2.6BFP85.6 GB17 sRuns well
50Fair quality (C) · 50 of 100: how capable the model is at FP8, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.Stable CascadeStability AI · 5.2BFP88.1 GB19 sRuns well
52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.Hunyuan-DiT v1.2Tencent · 1.5BFP169.3 GB25 sRuns well
40Fair quality (C) · 40 of 100: how capable the model is at FP16, whatever it runs on. Match score 60/100 (quality, fit, speed and memory use) sets the ranking.SDXL TurboStability AI · 2.6BFP167.6 GB0.6 sRuns great
42Fair quality (C) · 42 of 100: how capable the model is at FP16, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.Realistic Vision V5.1SG161222 · 0.9BFP162.8 GB5.1 sRuns great
40Fair quality (C) · 40 of 100: how capable the model is at FP16, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.DreamShaper 8Lykon · 0.9BFP162.8 GB5.1 sRuns great
35Basic quality (D) · 35 of 100: how capable the model is at FP16, whatever it runs on. Match score 56/100 (quality, fit, speed and memory use) sets the ranking.Stable Diffusion 1.5Stability AI / Runway · 0.9BFP162.8 GB5.1 sRuns great
30Basic quality (D) · 30 of 100: how capable the model is at FP16, whatever it runs on. Match score 54/100 (quality, fit, speed and memory use) sets the ranking.LCM DreamShaper v7Tsinghua (LCM) · 0.9BFP162.8 GB1.1 sRuns great
28Basic quality (D) · 28 of 100: how capable the model is at FP16, whatever it runs on. Match score 53/100 (quality, fit, speed and memory use) sets the ranking.SD TurboStability AI · 0.9BFP163.3 GB0.3 sRuns great

Video models on the Apple M3 Ultra (96 GB)

Quality Model Quant Memory Speed Runs
86Excellent quality (A+) · 86 of 100: how capable the model is at FP8, whatever it runs on. Match score 89/100 (quality, fit, speed and memory use) sets the ranking.LTX-2.5 (22B, video + audio)Lightricks · 21.0BFP848.3 GB5.8 minRuns well
84Great quality (A) · 84 of 100: how capable the model is at FP8, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.LTX-2.3Lightricks · 18.9BFP844.3 GB25.4 minSlow
85Excellent quality (A+) · 85 of 100: how capable the model is at FP16, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.Bernini-R A14B (video generation + editing)ByteDance · 14.3BFP1646.0 GB50.8 minVery slow
86Excellent quality (A+) · 86 of 100: how capable the model is at FP16, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.Wan 2.2 I2V A14BAlibaba · 14.0BFP1658.0 GB1.9 hVery slow
86Excellent quality (A+) · 86 of 100: how capable the model is at FP16, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.Wan 2.2 T2V A14BAlibaba · 14.0BFP1658.0 GB1.9 hVery slow
89Excellent quality (A+) · 89 of 100: how capable the model is at Q8 GGUF, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.MiniMax H3 (33B, video + audio)MiniMax · 20.1BQ8 GGUF68.2 GB84.3 minVery slowtight fit
80Great quality (A) · 80 of 100: how capable the model is at FP8, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.LTX-2 (19B, video + audio)Lightricks · 18.9BFP844.3 GB25.4 minSlow
84Great quality (A) · 84 of 100: how capable the model is at FP16, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.HunyuanVideo 1.5 (8.3B)Tencent · 8.3BFP1655.2 GB2.1 hVery slow
80Great quality (A) · 80 of 100: how capable the model is at FP16, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.SCAIL-2 14B (character animation)Zhipu AI · 16.4BFP1652.5 GB84.6 minVery slow
74Great quality (A) · 74 of 100: how capable the model is at FP16, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.FramePack I2V (HunyuanVideo 13B)lllyasviel · 12.9BFP1649.6 GB28.2 minSlow
66Good quality (B) · 66 of 100: how capable the model is at FP8, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.LTX-Video 13B 0.9.8 (distilled)Lightricks · 13.0BFP830.1 GB3.8 minRuns great
76Great quality (A) · 76 of 100: how capable the model is at FP16, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Wan 2.2 S2V 14B (speech to video)Alibaba · 16.3BFP1650.4 GB70.5 minVery slow
78Great quality (A) · 78 of 100: how capable the model is at FP16, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Wan 2.1 I2V 14B 720PAlibaba · 14.0BFP1659.3 GB2.4 hVery slow
78Great quality (A) · 78 of 100: how capable the model is at FP16, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Wan 2.1 T2V 14BAlibaba · 14.0BFP1658.0 GB2.4 hVery slow
78Great quality (A) · 78 of 100: how capable the model is at FP16, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.SkyReels V2 T2V 14B 720PSkywork · 14.0BFP1660.8 GB2.8 hVery slow
76Great quality (A) · 76 of 100: how capable the model is at FP16, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.HY-OmniWeaving (8.3B, unified video)Tencent · 8.3BFP1641.9 GB37.4 minSlow
74Great quality (A) · 74 of 100: how capable the model is at FP16, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking.Wan 2.1 I2V 14B 480PAlibaba · 14.0BFP1646.5 GB63.5 minVery slow
74Great quality (A) · 74 of 100: how capable the model is at FP16, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.SkyReels V2 T2V 14B 540PSkywork · 14.0BFP1649.0 GB1.6 hVery slow
77Great quality (A) · 77 of 100: how capable the model is at Q8 GGUF, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Wan 2.2 Animate 14BAlibaba · 17.3BQ8 GGUF42.7 GB2.6 hVery slow
77Great quality (A) · 77 of 100: how capable the model is at Q8 GGUF, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Wan 2.1 VACE 14BAlibaba · 17.3BQ8 GGUF42.8 GB3.0 hVery slow
74Great quality (A) · 74 of 100: how capable the model is at FP8, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Step-Video T2V (30B)StepFun · 30.0BFP846.8 GB3.3 hVery slow
74Great quality (A) · 74 of 100: how capable the model is at FP8, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Wan-Dancer 14B (music to dance)Alibaba · 17.3BFP853.4 GB4.7 hVery slow
70Great quality (A) · 70 of 100: how capable the model is at FP16, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Wan 2.2 TI2V 5BAlibaba · 5.0BFP1643.1 GB36.7 minSlow
88Excellent quality (A+) · 88 of 100: how capable the model is at FP16, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.Kandinsky 6.0 Pro (29B, video + audio)Sber AI · 30.1BFP1668.6 GB3.6 hVery slowtight fit · encoder unloaded after use
70Great quality (A) · 70 of 100: how capable the model is at FP16, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Cosmos-Predict2 14B Video2WorldNVIDIA · 14.0BFP1657.6 GB1.9 hVery slow
71Great quality (A) · 71 of 100: how capable the model is at Q8 GGUF, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.HunyuanVideo 13BTencent · 13.0BQ8 GGUF44.6 GB3.2 hVery slow
71Great quality (A) · 71 of 100: how capable the model is at Q8 GGUF, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.HunyuanVideo I2V 13BTencent · 13.0BQ8 GGUF44.6 GB3.2 hVery slow
70Great quality (A) · 70 of 100: how capable the model is at FP16, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Open-Sora 2.0 (11B)HPC-AI Tech · 11.0BFP1643.8 GB3.5 hVery slow
68Good quality (B) · 68 of 100: how capable the model is at FP16, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.Kandinsky 6.0 Video Lite (3B, video + audio)Sber AI · 3.2BFP1631.5 GB21.9 minSlow
60Good quality (B) · 60 of 100: how capable the model is at FP16, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.Cosmos 1.0 Diffusion 7B Text2WorldNVIDIA · 7.0BFP1648.2 GB38.8 minSlow
60Good quality (B) · 60 of 100: how capable the model is at FP16, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking.CogVideoX 1.5 5BZhipu AI · 5.6BFP1650.3 GB63.5 minVery slow
62Good quality (B) · 62 of 100: how capable the model is at FP16, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.Mochi 1 Preview (10B)Genmo · 10.0BFP1641.4 GB77.6 minVery slow
88Excellent quality (A+) · 88 of 100: how capable the model is at FP16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.Cosmos3-Super Image2Video (64B)NVIDIA · 64.0BFP1653.7 GB3.4 hVery slowCPU offload
50Fair quality (C) · 50 of 100: how capable the model is at FP16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.LTX-Video 2BLightricks · 1.9BFP1617.5 GB1.8 minRuns great
58Good quality (B) · 58 of 100: how capable the model is at FP16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking.Cosmos-Predict2 2B Video2WorldNVIDIA · 2.0BFP1633.6 GB17.6 minSlow
52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.Allegro (2.8B)Rhymes AI · 2.8BFP1638.2 GB84.6 minVery slow
55Good quality (B) · 55 of 100: how capable the model is at FP16, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Wan 2.1 VACE 1.3BAlibaba · 2.2BFP1621.5 GB18.3 minSlow
55Good quality (B) · 55 of 100: how capable the model is at FP16, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking.Wan 2.1 T2V 1.3BAlibaba · 1.3BFP1619.8 GB16.9 minSlow
53Fair quality (C) · 53 of 100: how capable the model is at FP8, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.CogVideoX 5B I2VZhipu AI · 5.6BFP814.9 GB11.4 minRuns well
52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.SkyReels V2 Diffusion Forcing 1.3BSkywork · 1.4BFP1623.8 GB21.2 minSlow
44Fair quality (C) · 44 of 100: how capable the model is at FP8, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.CogVideoX 2BZhipu AI · 1.7BFP811.0 GB3.8 minRuns great
53Fair quality (C) · 53 of 100: how capable the model is at FP8, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.CogVideoX 5BZhipu AI · 5.0BFP814.3 GB11.4 minRuns well
35Basic quality (D) · 35 of 100: how capable the model is at FP16, whatever it runs on. Match score 57/100 (quality, fit, speed and memory use) sets the ranking.AnimateDiff v3 (SD 1.5)Shanghai AI Lab · 1.3BFP164.7 GB85 sRuns great

Decision models on the Apple M3 Ultra (96 GB)

Quality Model Quant Memory Speed Runs
86Excellent quality (A+) · 86 of 100: how capable the model is at F16, whatever it runs on. Match score 93/100 (quality, fit, speed and memory use) sets the ranking.Jina Reranker v3.5Jina AI · 600MF167.8 GB24 msRuns great
90Excellent quality (A+) · 90 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 93/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 Reranker 4BAlibaba · 4.0BQ8_010.9 GB130 msRuns great
84Great quality (A) · 84 of 100: how capable the model is at F16, whatever it runs on. Match score 92/100 (quality, fit, speed and memory use) sets the ranking.Intern-Decision 4BShanghai AI Lab · 4.5BF1615.5 GB92 msRuns great
84Great quality (A) · 84 of 100: how capable the model is at F16, whatever it runs on. Match score 92/100 (quality, fit, speed and memory use) sets the ranking.Jina Reranker v3Jina AI · 600MF167.8 GB24 msRuns great
88Excellent quality (A+) · 88 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 92/100 (quality, fit, speed and memory use) sets the ranking.zerank-2ZeroEntropy · 4.0BQ8_010.9 GB130 msRuns great
86Excellent quality (A+) · 86 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 92/100 (quality, fit, speed and memory use) sets the ranking.Shieldstral 1.0 3BMistral AI · 3.9BQ8_010.7 GB125 msRuns great
82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking.d1 3BLiquid AI · 3.1BF1612.6 GB65 msRuns great
82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking.mxbai-rerank-large v2Mixedbread · 1.5BF169.5 GB56 msRuns great
80Great quality (A) · 80 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking.FRIDA-DecisionsSber AI · 820MF162.4 GB19 msRuns great
80Great quality (A) · 80 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking.GLiNER2.5 DecideFastino · 490MF161.7 GB13 msRuns great
80Great quality (A) · 80 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking.LemonDecider 4BAnt Group (inclusionAI) · 4.5BF1610.8 GB87 msRuns great
80Great quality (A) · 80 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 Reranker 0.6BAlibaba · 600MF167.8 GB24 msRuns great
80Great quality (A) · 80 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking.zerank-1 smallZeroEntropy · 1.7BF1610.0 GB62 msRuns great
78Great quality (A) · 78 of 100: how capable the model is at F16, whatever it runs on. Match score 90/100 (quality, fit, speed and memory use) sets the ranking.BGE Reranker v2 GemmaBAAI · 2.5BF1611.4 GB89 msRuns great
90Excellent quality (A+) · 90 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 89/100 (quality, fit, speed and memory use) sets the ranking.Qwen3 Reranker 8BAlibaba · 8.2BQ8_015.3 GB261 msRuns well
76Great quality (A) · 76 of 100: how capable the model is at F16, whatever it runs on. Match score 89/100 (quality, fit, speed and memory use) sets the ranking.Llama Prompt Guard 2 86MMeta · 280MF161.2 GB14 msRuns great
76Great quality (A) · 76 of 100: how capable the model is at F16, whatever it runs on. Match score 89/100 (quality, fit, speed and memory use) sets the ranking.mxbai-rerank-base v2Mixedbread · 490MF167.4 GB21 msRuns great
74Great quality (A) · 74 of 100: how capable the model is at F16, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking.GLiClass ModernBERT Large v3Knowledgator · 400MF161.5 GB11 msRuns great
78Great quality (A) · 78 of 100: how capable the model is at F16, whatever it runs on. Match score 88/100 (quality, fit, speed and memory use) sets the ranking.OpenAI Privacy FilterOpenAI · 1.4B (50M active)F163.8 GB4.9 msRuns great
72Great quality (A) · 72 of 100: how capable the model is at F16, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.BGE Reranker v2 M3BAAI · 570MF161.9 GB25 msRuns great
72Great quality (A) · 72 of 100: how capable the model is at F16, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.DeBERTa v3 Large Zero-shot v2.0Moritz Laurer · 440MF161.6 GB12 msRuns great
72Great quality (A) · 72 of 100: how capable the model is at F16, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.DuoGuard 0.5BDuoGuard · 490MF161.7 GB22 msRuns great
72Great quality (A) · 72 of 100: how capable the model is at F16, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.Skywork Reward V2 Qwen3 0.6BSkywork · 600MF162.0 GB48 msRuns great
90Excellent quality (A+) · 90 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 87/100 (quality, fit, speed and memory use) sets the ranking.Skywork Reward V2 Llama 3.1 8BSkywork · 7.5BQ8_010.4 GB478 msSlow
70Great quality (A) · 70 of 100: how capable the model is at F16, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.GTE Reranker ModernBERT BaseAlibaba · 150MF160.9 GB9.5 msRuns great
70Great quality (A) · 70 of 100: how capable the model is at F16, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.Jina Reranker v2 Base MultilingualJina AI · 280MF161.2 GB14 msRuns great
70Great quality (A) · 70 of 100: how capable the model is at F16, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.ModernBERT Large Zero-shot v2.0Moritz Laurer · 400MF161.5 GB11 msRuns great
86Excellent quality (A+) · 86 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking.GLiDE 27BFastino · 26BQ8_034.8 GB416 msSlow
83Great quality (A) · 83 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.Qwen3Guard Gen 4BAlibaba · 4.4BQ6_K10.3 GB176 msRuns well
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.GLiGuard 300MFastino · 210MF161.1 GB12 msRuns great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.Granite Embedding Reranker English R2IBM · 150MF160.9 GB9.5 msRuns great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.LayaConvai Innovations · 420MF161.5 GB12 msRuns great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.Llama Prompt Guard 2 22MMeta · 71MF160.8 GB6.6 msRuns great
74Great quality (A) · 74 of 100: how capable the model is at F16, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking.Qwen3Guard Gen 0.6BAlibaba · 750MF168.1 GB48 msRuns great
66Good quality (B) · 66 of 100: how capable the model is at F16, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking.GLiClass Base v3Knowledgator · 190MF161.0 GB7.5 msRuns great
66Good quality (B) · 66 of 100: how capable the model is at F16, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking.ProtectAI DeBERTa v3 Prompt Injection v2Protect AI · 180MF161.0 GB11 msRuns great
84Great quality (A) · 84 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking.Skywork Reward V2 Qwen3 8BSkywork · 7.6BQ8_010.5 GB483 msSlow
64Good quality (B) · 64 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.DeBERTa v3 Base Zero-shot v2.0Moritz Laurer · 180MF161.0 GB7.3 msRuns great
86Excellent quality (A+) · 86 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.Qwen3Guard Gen 8BAlibaba · 8.2BQ8_015.3 GB354 msRuns well
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.GLiNER2 Base v1Fastino · 210MF161.1 GB7.8 msRuns great
66Good quality (B) · 66 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.NuExtract 2.0 2BNuMind · 2.2BQ4_K_M7.8 GB97 msRuns great
62Good quality (B) · 62 of 100: how capable the model is at F16, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.HHEM 2.1 OpenVectara · 110MF160.8 GB12 msRuns great
84Great quality (A) · 84 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.Llama 3.1 Nemotron Safety Guard 8B v3NVIDIA · 8.0BQ8_015.1 GB347 msRuns well
66Good quality (B) · 66 of 100: how capable the model is at F16, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.LettuceDetect LargeKR Labs · 400MF161.5 GB33 msRuns great
84Great quality (A) · 84 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.Granite Guardian 4.1 8BIBM · 8.4BQ8_015.6 GB362 msRuns well
72Great quality (A) · 72 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.NuExtract 2.0 4BNuMind · 3.8BQ4_K_M8.7 GB162 msRuns well
60Good quality (B) · 60 of 100: how capable the model is at F16, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking.Laya MultilingualConvai Innovations · 320MF161.3 GB9.8 msRuns great
79Great quality (A) · 79 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.WildGuard 7BAi2 · 7.3BQ6_K12.6 GB287 msRuns well
58Good quality (B) · 58 of 100: how capable the model is at F16, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking.ShieldGemma 2BGoogle · 2.6BF1611.8 GB93 msRuns great
80Great quality (A) · 80 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.Granite Guardian 3.3 8BIBM · 8.2BQ8_015.4 GB353 msRuns well
62Good quality (B) · 62 of 100: how capable the model is at F16, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.GLiNER Multi v2.1Urchade Zaratiana · 290MF161.3 GB9.3 msRuns great
56Good quality (B) · 56 of 100: how capable the model is at F16, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.BART Large MNLIMeta · 410MF161.5 GB11 msRuns great
52Fair quality (C) · 52 of 100: how capable the model is at F16, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking.ms-marco MiniLM L6 v2Sentence Transformers · 23MF160.7 GB4.8 msRuns great
64Good quality (B) · 64 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.ShieldGemma 9BGoogle · 9.2BQ8_016.9 GB294 msRuns well
56Good quality (B) · 56 of 100: how capable the model is at F16, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking.Llama Guard 3 1BMeta · 1.5BF169.4 GB92 msRuns great
67Good quality (B) · 67 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking.Llama Guard 3 8BMeta · 8.0BQ6_K13.2 GB318 msRuns well
66Good quality (B) · 66 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.ShieldGemma 27BGoogle · 27BQ8_036.0 GB857 msSlow
86Excellent quality (A+) · 86 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking.CompassJudger-2 32BOpenCompass · 33BQ8_041.6 GB13.1 sVery slow
77Great quality (A) · 77 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Nemotron Content Safety Reasoning 4BNVIDIA · 4.3BQ6_K10.2 GB1.3 sSlow
76Great quality (A) · 76 of 100: how capable the model is at MXFP4, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.gpt-oss-safeguard 20BOpenAI · 22B (3.6B active)MXFP418.7 GB1.3 sSlow
65Good quality (B) · 65 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.Llama Guard 4 12BMeta · 12BQ6_K16.6 GB473 msSlow
78Great quality (A) · 78 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking.DynaGuard 8BUniversity of Maryland · 8.2BQ6_K13.4 GB2.4 sVery slow
79Great quality (A) · 79 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.Selene 1 Mini 8BAtla · 8.0BQ6_K13.2 GB2.5 sVery slow
69Good quality (B) · 69 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking.DynaGuard 1.7BUniversity of Maryland · 2.0BQ6_K8.3 GB604 msSlow
77Great quality (A) · 77 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking.CompassJudger-2 7BOpenCompass · 7.6BQ6_K12.7 GB2.4 sVery slow
74Great quality (A) · 74 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking.Prometheus 2 8x7BPrometheus Eval · 47B (13B active)Q8_056.1 GB5.7 sVery slow
65Good quality (B) · 65 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.GLIDERPatronus AI · 3.8BQ6_K10.3 GB1.2 sSlow
67Good quality (B) · 67 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 67/100 (quality, fit, speed and memory use) sets the ranking.Prometheus 2 7BPrometheus Eval · 7.2BQ6_K12.6 GB2.2 sVery slow
63Good quality (B) · 63 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 66/100 (quality, fit, speed and memory use) sets the ranking.Flow Judge v0.1Flow AI · 3.8BQ6_K10.3 GB1.2 sSlow
80Great quality (A) · 80 of 100: how capable the model is at MXFP4, whatever it runs on. Match score 59/100 (quality, fit, speed and memory use) sets the ranking.gpt-oss-safeguard 120BOpenAI · 117B (5.1B active)MXFP473.0 GB1.8 sVery slowover GPU memory limit

Upgrade picks

GPUs that would run more LLMs without offloading, at launch prices. Right now 525 of 622 LLMs run well on your Apple M3 Ultra (96 GB).

NVIDIA RTX PRO 6000 Blackwell Workstation Edition Cheapest upgrade NVIDIA RTX PRO 6000 Blackwell 96.0 GB · $8,565 launch price +22 LLMs run well, including Inkling Small (276B-A12B)

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.

Was this page helpful?