Mellum2.1 Thinking (12B-A2.5B) on Intel Xeon D-1600 series NAS (8 GB)

Can the Intel Xeon D-1600 series NAS (8 GB) run Mellum2.1 Thinking (12B-A2.5B)?

Yes, but slowly (far over memory limit)

Slow (far over memory limit). At Q2_K it needs about 6.0 GB of 5.2 GB usable, with an estimated speed of 15.5 tok/s.

Best quant: Q2_K
Generation speed: 15.5 tok/s
Prompt processing: 45.0 tok/s
First token (1K prompt): 22.8 s
Largest safe context: —

Best app to run it on the Intel Xeon D-1600 series NAS (8 GB)

On Linux

llama.cpp

Command line + local API · runs on llama.cpp

Splits the model between GPU memory and system RAM, which this setup needs.
Gives you direct control of how many layers sit on the GPU, the best way to squeeze a model that doesn’t quite fit.
Usually the first to support new model architectures.

llama-server -hf JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF:Q2_K

Ollama Splits the model between GPU memory and system RAM, which this setup needs.
KoboldCpp Splits the model between GPU memory and system RAM, which this setup needs.

Want an agent? Mellum2.1 Thinking (12B-A2.5B) 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, which apps run in a container on a NAS without a graphics card (models run on its processor), and whether the model fits fully in the memory models can use. Details.

Download Mellum2.1 Thinking (12B-A2.5B)

Quantized files · GGUF

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

huggingface.co/JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF (opens in a new tab)

ollama run hf.co/JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF

Original weights · Hugging Face

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

huggingface.co/JetBrains/Mellum2.1-12B-A2.5B-Thinking (opens in a new tab)

No sign-in needed.

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

How it’s rated

Quality 49Fair quality (C) · 49 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 50/100 (quality, fit, speed and memory use) sets the ranking.: How capable the model itself is at Q2_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 Slowfar over memory limit: How well it runs on the Intel Xeon D-1600 series NAS (8 GB): 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.	22.1 / 45
Memory fitFull marks when the model fits in the share of the NAS’s memory that models can use (the rest stays with the NAS system and its storage services); fewer when it goes past that limit.	4.0 / 25
SpeedGeneration speed against a 40 tokens/s target for LLMs, or render time against a target for image and video.	9.3 / 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	50.4 / 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 Intel Xeon D-1600 series NAS (8 GB)

Quant	Memory needed	Available	Speed	Runs	Quality	Download
`Q2_K`	6.0 GB	5.2 GB	15.5 tok/s	Slowfar over memory limit	49Fair quality (C) · 49 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 50/100 (quality, fit, speed and memory use) sets the ranking.	GGUF files (opens in a new tab)
`Q3_K_M`	7.1 GB	5.2 GB	—	Won't run	58Good quality (B) · 58 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 7/100 (quality, fit, speed and memory use) sets the ranking.	GGUF files (opens in a new tab)
`Q4_K_M`	8.5 GB	5.2 GB	—	Won't run	66Good quality (B) · 66 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 8/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q5_K_M`	9.8 GB	5.2 GB	—	Won't run	67Good quality (B) · 67 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.	GGUF files (opens in a new tab)
`Q6_K`	11.1 GB	5.2 GB	—	Won't run	69Good quality (B) · 69 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q8_0`	14.0 GB	5.2 GB	—	Won't run	70Great quality (A) · 70 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`F16`	25.4 GB	5.2 GB	—	Won't run	70Great quality (A) · 70 of 100: how capable the model is at F16, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab) 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

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. NVIDIA GeForce RTX 5090 F16 · 108 tok/s · Runs great
69Good quality (B) · 69 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking. AMD Radeon RX 9070 XT Q6_K · 112 tok/s · Runs great
69Good quality (B) · 69 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking. Apple M4 Pro (24 GB) Q6_K · 62.1 tok/s · Runs great
69Good quality (B) · 69 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 4060 Ti 16GB Q6_K · 72.5 tok/s · Runs great
69Good quality (B) · 69 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 4070 Ti SUPER Q6_K · 125 tok/s · Runs great
69Good quality (B) · 69 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 4080 SUPER Q6_K · 132 tok/s · Runs great

Quick answers

How much memory does Mellum2.1 Thinking (12B-A2.5B) need on the Intel Xeon D-1600 series NAS 8GB?

At Q2_K it needs about 6.0 GB. The Intel Xeon D-1600 series NAS 8GB has 5.2 GB usable.

How fast is Mellum2.1 Thinking (12B-A2.5B) on the Intel Xeon D-1600 series NAS 8GB?

About 15.5 tok/s at Q2_K, an estimate from the hardware’s memory bandwidth and the size of the model.

Which quantization of Mellum2.1 Thinking (12B-A2.5B) should I use on the Intel Xeon D-1600 series NAS 8GB?

Q2_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.