Can I run this LLM?
Can a Mac Studio M3 Ultra (512GB) run Devstral Small 2 24B?
Mistral · See model card · estimated for a 512.0 GB machine at 4K context.
Yes — it runs great.
Runs greatBest fit: Devstral Small 2 24B at Q8_0 · 25.1 GB weights · 19–29 tok/s on a Mac Studio M3 Ultra (512GB).
Arena Text Overall · style-controlled
UnratedNo exact matchNo reviewed exact Arena match
No reviewed exact model/configuration match in the Arena Text Overall style-controlled snapshot. No family or editorial score is substituted. Local quantization quality is unmeasured.
Snapshot published 2026-10-08 · snapshot imported 2026-10-11 UTC · validated last-good data · manual refresh
Arena leaderboard dataset · Arena / lmarena-ai, Arena Leaderboard Dataset (CC BY 4.0). Selected Text Overall style-controlled rows; field names adapted, ratings unchanged. · dataset CC BY 4.0 · revision e16f90232986
- Total memory at 4K context
- 26.5 GB
- Weights + KV cache + runtime overhead
- Estimated generation
- 19–29 tok/s
- Estimated first token
- ~10s
Every quantization on this machine
On small screens, each quant lists every metric below its name. On wider screens, scroll sideways for total memory, speed and fit. Keyboard: focus the table, then use the arrow keys. Quant names stay visible.
| Quant | Weights | Total memory | Speed | Fit |
|---|---|---|---|---|
| Q2_K | 8.9 GB | 10.4 GB | 43–65 tok/s | Runs great |
| Q3_K_M | 11.5 GB | 12.9 GB | 36–54 tok/s | Runs great |
| Q4_K_M | 14.3 GB | 15.8 GB | 30–45 tok/s | Runs great |
| Q5_K_M | 16.8 GB | 18.2 GB | 27–40 tok/s | Runs great |
| Q6_K | 19.4 GB | 20.8 GB | 24–36 tok/s | Runs great |
| Q8_0 | 25.1 GB | 26.5 GB | 19–29 tok/s | Runs great |
| FP16 | 47.2 GB | 48.6 GB | 11–16 tok/s | Tight / slow |
Ran Devstral Small 2 24B on a Mac Studio M3 Ultra (512GB)? Add your real numbers →