Yes, just — 4.3 GB spare

Mistral Small 4 119B-A6B at Q4_K_M fits your M2 Max · 96 GB entirely in unified memory at 8K context, at an estimated 29 tokens per second. Past 204K the KV cache pushes it over — quantise the cache to q8_0, or step down a quantisation, to go longer.

Fully in unified memory 8K context Q4_K_M · 66.9 GB Apache 2.0 Released 17 Mar 2026 Vision Not in the Ollama library

Instruct, reasoning, vision and code in one 119B MoE. A latent KV cache (320 wide) keeps context cheap; the weights still want 64 GB+.

What hardware do I need for Mistral Small 4 119B-A6B? →

The VRAM budget

weights 66.9 GB

Weights 66.9 GB KV cache @ 8K 0.18 GB Runtime overhead 0.6 GB Free 4.3 GB of 72.0 GB

Quantisation ladder

Quant	Weights	Total @ 8K	Max context	Tok/s	Quality	Fit
Q8_0	117.8 GB	118.5 GB	—	~16	−0.1% ppl	46.5 GB over
Q6_K	90.9 GB	91.7 GB	—	~21	−0.4% ppl	19.7 GB over
Q5_K_M	78.5 GB	79.3 GB	—	~24	−0.8% ppl	7.3 GB over
Q4_K_M	66.9 GB	67.7 GB	204K	29	−1.9% ppl	Recommended
Q3_K_M	54.2 GB	54.9 GB	256K	35	−5.4% ppl	Long context
Q2_K	46.4 GB	47.2 GB	256K	41	−15% ppl	Long context

Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it. This model uses multi-head latent attention, so its cache is a compressed latent rather than full K and V.

How to run it

llama.cpp LM Studio MLX

terminal

$ pip install mlx-lm
$ mlx_lm.generate --model mlx-community/Mistral-Small-4-119B-2603-4bit \
    --max-tokens 512 --prompt "Hello"

Apple's own array framework. The fastest path on Apple Silicon. More on MLX.

01Download is 66.9 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.

02macOS caps what the GPU may wire down at about 72.0 GB of 96 GB. Raising it with iogpu.wired_limit_mb is possible, and risky.

03Only 4.3 GB is spare, so a long system prompt can still push it over. Its real ceiling here is 204K context.

See all models for this rig Compare with another model