Can I Run Devstral Small 2505 24B on Apple M5 Max?
Written by Jakub Rusinowski · Last updated September 29, 2026
Yes. Devstral Small 2505 24B needs 16.6 GB at 8K context (14.5 GB weights + 2.1 GB KV cache/overhead) at Q4_K_M and the Apple M5 Max has 96 GB usable, leaving about 79.4 GB of headroom, at roughly 21 tok/s (estimated) with room for up to 128K of context.
See what else this hardware can run →
The numbers
| Model | Devstral Small 2505 24B |
| Parameters | 24B |
| Quantization | Q4_K_M |
| VRAM needed (8K context) | 16.6 GB |
| Apple M5 Max usable memory | 96 GB |
| Verdict | Runs |
| Max context | 128K |
| Estimated speed | ~21 tok/s |
Devstral Small 2505 24B on the Apple M5 Max: VRAM by quantization
| Quant | VRAM needed | Fits 128 GB? | Max context | Whole PC |
|---|---|---|---|---|
| F16 | 50.1 GB | ✓ Yes | 128K | — |
| Q8_0 | 27.6 GB | ✓ Yes | 128K | — |
| Q6_K | 21.8 GB | ✓ Yes | 128K | — |
| Q5_K_M | 19.2 GB | ✓ Yes | 128K | — |
| Q4_K_M | 16.6 GB | ✓ Yes | 128K | — |
| Q3_K_M | 12.4 GB | ✓ Yes | 128K | — |
| Q2_K | 10 GB | ✓ Yes | 128K | — |
Assumes an 8K-token context with an f16 KV cache. A longer window needs more; a quantized KV cache needs less. “Max context” is the largest window that still fits in 128 GB of unified memory. Figures are estimates from parameter count, quantization and memory bandwidth — the analyzer lets you tune KV-cache quant and context.
Context costs VRAM too. At Q4_K_M on the Apple M5 Max: 8K 16.6 GB · 32K 20.7 GB · 128K 36.8 GB. Every window this model publishes still fits 128 GB.
What to do next
Nothing to change — run it at Q4_K_M. If you want more capability, check whether the next size up also fits.
No compatible GPU? Devstral Small 2505 24B on 16 GB of system RAM, CPU only: runs, but it is tight.
Running Devstral Small 2505 24B on this card:
Devstral Small 2505 24B needs ~17 GB but this GPU has 128 GB. Rent a RTX 4090 (24 GB)-class GPU by the hour instead of buying one:
Affiliate links — we may earn a commission if you sign up, at no extra cost to you.
Cloud rates verified 2026-07 — estimates, and marketplace prices vary. Buying price is GPU MSRP only, not a full PC.
Other Devstral sizes on the Apple M5 Max
- Devstral-2 123B — runs (77.9 GB)
- Devstral Small 2 24B — runs (16.6 GB)
Devstral Small 2505 24B on nearby hardware
- NVIDIA RTX PRO 6000 Blackwell (96 GB) — runs
All Devstral sizes on the Apple M5 Max | Devstral Small 2505 24B model page | Apple M5 Max GPU page | VRAM calculator | Check your hardware