Mixtral 8x7B on Apple M2 Ultra · 192GB

✅ Runs comfortably — Mixtral 8x7B @ Q8_0
Mixtral 8x7B 8x7B (MoE) at Q8_0 needs ~49.5 GB (weights 46.2 GB + KV 1.0 GB + overhead 2.3 GB @ 8K ctx) of your 141 GB usable (of 192 GB unified memory) — plenty of headroom. Expect ~17–30 tok/s (fast).
context @ Q8_0: comfortable to 32K
quantneedsspeed
✓ Q8_049.5 GB~17–30 tok/s (fast)
✓ Q5_K_M33.5 GB~25–40 tok/s (fast)
✓ Q4_K_M28.7 GB~30–45 tok/s (fast)
✓ Q3_K_M23.3 GB~35–55 tok/s (fast)
$runlocal install mixtral:8x7b
runlocal verdict card
download card share on x share on reddit drop it in your model-drop post
Mixtral 8x7B: Runs comfortably
for your own README — links back here

Same model, other machines: Apple M1 · Apple M1 Pro · Apple M2 · Apple M3 Pro · Apple M2 Pro · Apple M4 Pro

Also runs on a Apple M2 Ultra: DeepSeek-R1 (distill) · Qwen3 · Qwen2.5 · Llama 3.1 · Llama 3.3

Best models for a Apple M2 Ultra · 192GB →

Architecturally similar

These share Mixtral 8x7B's KV-cache geometry — they size memory the same way, so here's how they fit on a Apple M2 Ultra · 192GB too:

modelsharesverdict here
Qwen2.5-Coder 8kv/128hd ✅ Runs comfortably
Llama 3.2 8kv/128hd ✅ Runs comfortably
Mistral 7B 8kv/128hd ✅ Runs comfortably
Mistral Small 3 8kv/128hd ✅ Runs comfortably
Mistral Nemo 8kv/128hd ✅ Runs comfortably
Codestral 8kv/128hd ✅ Runs comfortably
Llama 3.2 Vision 8kv/128hd ✅ Runs comfortably
Phi-4-mini 8kv/128hd ✅ Runs comfortably

…and 6 more — see the Mixtral 8x7B page.

Check any combo yourself: open the checker →