SmolLM2 on Apple M3 Max · 36GB

✅ Runs comfortably — SmolLM2 1.7B @ FP16
SmolLM2 1.7B at FP16 needs ~5.2 GB (weights 3.2 GB + KV 1.5 GB + overhead 512 MB @ 8K ctx) of your 21.1 GB usable (of 36.0 GB unified memory) — plenty of headroom. Expect ~40–65 tok/s (very fast).
context @ FP16: comfortable to 8K
quantneedsspeed
✓ FP165.2 GB~40–65 tok/s (very fast)
✓ Q8_03.7 GB~55–90 tok/s (very fast)
✓ Q4_K_M3.0 GB~70–120 tok/s (very fast)
$runlocal install smollm2:1.7b
runlocal verdict card
download card share on x share on reddit drop it in your model-drop post
SmolLM2: 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 M3 Max: Qwen3 · DeepSeek-R1 (distill) · Qwen2.5 · Gemma 4 · Gemma 3

Best models for a Apple M3 Max · 36GB →

Check any combo yourself: open the checker →