Can the AMD Instinct MI100 run Sage-A1.5 35B-A3B?

Yes, it runs well (tight fit)

Runs well (tight fit). At Q6_K it needs about 31.8 GB of 32.0 GB usable, with an estimated speed of 144 tok/s.

Best quant: Q6_K
Generation speed: 144 tok/s
Prompt processing: 9,527 tok/s
First token (1K prompt): 0.3 s
Largest safe context: 15K

Best app to run it on the AMD Instinct MI100

On Linux

Lemonade

Desktop app · runs on llama.cpp, ONNX Runtime GenAI (Ryzen AI NPU)

AMD’s own app: llama.cpp tuned for Radeon on Vulkan and ROCm.

lemonade pull mradermacher/Sage-A1.5-i1-GGUF:Q6_K

LM Studio Point-and-click: search, download and chat in one window.
llama.cpp Usually the first to support new model architectures.

On Windows

Lemonade

Desktop app · runs on llama.cpp, ONNX Runtime GenAI (Ryzen AI NPU)

AMD’s own app: llama.cpp tuned for Radeon on Vulkan and ROCm.

lemonade pull mradermacher/Sage-A1.5-i1-GGUF:Q6_K

LM Studio Point-and-click: search, download and chat in one window.
llama.cpp Usually the first to support new model architectures.

Want an agent? Sage-A1.5 35B-A3B supports tool use, so you can pair a llama.cpp or MLX LM server with one of these. Apps that run on this computer come first.

Coding agents

Personal assistants

How we choose: the files published for this model, your operating system and GPU, whether it fits fully in memory, and published speed tests (MLX vs llama.cpp on Apple silicon, vLLM vs llama.cpp vs Ollama on NVIDIA). Details.

Download Sage-A1.5 35B-A3B

Recommended for the AMD Instinct MI100: Q6_K. Every other quantization has its own download link in the table below.

Quantized files · GGUF

Works in LM Studio, llama.cpp, Jan and Ollama. No sign-in needed.

Download the Q6_K file (opens in a new tab)

huggingface.co/mradermacher/Sage-A1.5-i1-GGUF (opens in a new tab)

ollama run hf.co/mradermacher/Sage-A1.5-i1-GGUF:Q6_K

Original weights · Hugging Face

Full-precision files from InternScience, for fine-tuning, vLLM, Transformers.

huggingface.co/InternScience/Sage-A1.5 (opens in a new tab)

No sign-in needed.

Official repo · GitHub

Reference code, inference scripts and docs from InternScience. The weights themselves download from Hugging Face.

github.com/InternScience/Agents-A1 (opens in a new tab)

Official repo linked from the Sage-A1.5 model card; its README currently documents Agents-A1 and the shared evaluation framework.

git clone https://github.com/InternScience/Agents-A1

Links checked 3 days ago against Hugging Face, Ollama and GitHub.

How it’s rated

Quality 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.: How capable the model itself is at Q6_K, from its editorial quality tier scaled by the quantization or precision. It’s the same on any hardware that runs this variant. Scale: A+ 85+ · A 70+ · B 55+ · C 40+ · D under 40.
Runs Runs welltight fit: How well it runs on the AMD Instinct MI100: speed, capped by how well it fits in memory. Runs great: 30+ tokens/s (or near the image, video or decision speed target) · Runs well: 10+ tokens/s · Slow: 3+ tokens/s · Very slow: under 3 · Won’t run: doesn’t fit even with offload. A tight fit or RAM offload caps it at Runs well; heavy offload caps it at Slow.

Match score

To order lists, each result also gets a match score out of 100 from four parts. It favors the most capable model that still runs well.

Model qualityEditorial quality tier of the model, scaled by the quantization or precision used.	40.5 / 45
Memory fitFull marks when everything sits in GPU memory; fewer when part spills to system RAM.	24.0 / 25
SpeedGeneration speed against a 40 tokens/s target for LLMs, or render time against a target for image and video.	15.0 / 15
Memory useRewards using the memory you have, so a bigger, better variant beats a tiny one that leaves most memory idle.	15.0 / 15
Match score	94.5 / 100

The match score orders results; it is not shown as a grade. Quality and Runs are the two ratings in the tables.

Every quantization on the AMD Instinct MI100

Quant	Memory needed	Available	Speed	Runs	Quality	Download
`Q2_K`	16.7 GB	32.0 GB	221 tok/s	Runs great	63Good quality (B) · 63 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q3_K_M`	20.0 GB	32.0 GB	199 tok/s	Runs great	76Great quality (A) · 76 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.	GGUF (opens in a new tab)
`Q4_K_M`	24.3 GB	32.0 GB	176 tok/s	Runs great	86Excellent quality (A+) · 86 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 93/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q5_K_M`	27.9 GB	32.0 GB	159 tok/s	Runs welltight fit	88Excellent quality (A+) · 88 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 93/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q6_K`	31.8 GB	32.0 GB	144 tok/s	Runs welltight fit	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.	GGUF (opens in a new tab)
`Q8_0`	40.5 GB	32.0 GB	42.8 tok/s	Runs welloffloads to RAM	91Excellent quality (A+) · 91 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.	GGUF files (opens in a new tab)
`F16`	74.3 GB	32.0 GB	—	Won't run	91Excellent quality (A+) · 91 of 100: how capable the model is at F16, whatever it runs on. Match score 12/100 (quality, fit, speed and memory use) sets the ranking.	Original weights (opens in a new tab)

Memory figures assume 8K of context. Longer conversations or documents grow the KV cache.

Other hardware that runs it

91Excellent quality (A+) · 91 of 100: how capable the model is at F16, whatever it runs on. Match score 95/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA RTX PRO 6000 Blackwell F16 · 98.2 tok/s · Runs great
90Excellent quality (A+) · 90 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 95/100 (quality, fit, speed and memory use) sets the ranking. Apple M1 Max (64 GB) Q6_K · 72.3 tok/s · Runs great
90Excellent quality (A+) · 90 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 95/100 (quality, fit, speed and memory use) sets the ranking. Apple M4 Max (64 GB) Q6_K · 90.1 tok/s · Runs great
88Excellent quality (A+) · 88 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 94/100 (quality, fit, speed and memory use) sets the ranking. Apple M4 Pro (48 GB) Q5_K_M · 62.2 tok/s · Runs 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. NVIDIA GeForce RTX 5090 Q6_K · 179 tok/s · Runs well (tight fit)
91Excellent quality (A+) · 91 of 100: how capable the model is at F16, whatever it runs on. Match score 94/100 (quality, fit, speed and memory use) sets the ranking. Apple M2 Ultra (192 GB) F16 · 51.3 tok/s · Runs great

Quick answers

How much memory does Sage-A1.5 35B-A3B need on the AMD Instinct MI100 32GB?

At Q6_K it needs about 31.8 GB. The AMD Instinct MI100 32GB has 32.0 GB usable.

How fast is Sage-A1.5 35B-A3B on the AMD Instinct MI100 32GB?

About 144 tok/s at Q6_K, an estimate from the hardware’s memory bandwidth and the size of the model.

Which quantization of Sage-A1.5 35B-A3B should I use on the AMD Instinct MI100 32GB?

Q6_K. Of the versions that fit, it is the one that runs best; when two run equally well, the check prefers the higher-rated one.

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.