Written by Jakub Rusinowski · Last updated August 15, 2026
Yes, but it is tight
Yes, but it is tight — Devstral Small 24B at Q6_K needs about 21.8 GB of the 24 GB usable on RTX 3090 Desktop (24 GB VRAM, 64 GB RAM), leaving only ~2.2 GB before the runtime starts swapping. Expect ~32 tok/s (estimated), with room for about 16,384 tokens of context.
Confidence: medium · Recommended quantization: Q6_K · Estimated speed: ~32 tok/s
| Usable memory for models | 24 GB |
| Memory bandwidth | 936 GB/s |
| Form factor | Desktop |
| Operating system | Windows or Linux |
| Memory upgradeable | Yes |
| Quant | Memory needed | Fits 24 GB? | Max context | Est. speed | Download |
|---|---|---|---|---|---|
| F16 | 50.1 GB | ✗ No | — | — | 48 GB |
| Q8_0 | 27.6 GB | ✗ No | — | — | 25.5 GB |
| Q6_K | 21.8 GB | ✓ Yes | 16K | ~32 tok/s | 19.7 GB |
| Q5_K_M | 19.2 GB | ✓ Yes | 32K | ~36.5 tok/s | 17 GB |
| Q4_K_M | 16.6 GB | ✓ Yes | 32K | ~41.9 tok/s | 14.5 GB |
| Q3_K_M | 12.4 GB | ✓ Yes | 64K | ~56.3 tok/s | 10.2 GB |
| Q2_K | 10 GB | ✓ Yes | 64K | ~69.2 tok/s | 7.9 GB |
| Variant | Needs at Q4_K_M | Fits? | Est. speed |
|---|---|---|---|
| Devstral-2 123B | 77.9 GB | ✗ Too large | — |
| Devstral Small 24B | 16.6 GB | ✓ Fits | ~41.9 tok/s |
| Devstral-2 22B | 15.7 GB | ✓ Fits | ~44.8 tok/s |
Ollama or llama.cpp (CUDA/ROCm) — vLLM if you need concurrent serving
Yes, but it is tight — Devstral Small 24B at Q6_K needs about 21.8 GB of the 24 GB usable on RTX 3090 Desktop (24 GB VRAM, 64 GB RAM), leaving only ~2.2 GB before the runtime starts swapping. Expect ~32 tok/s (estimated), with room for about 16,384 tokens of context.
Q6_K — it needs about 21.8 GB of the 24 GB available, downloads as roughly 19.7 GB, and runs at an estimated 32 tokens/sec with up to 16K of context.
Nothing binding — the model fits with headroom and generates at a usable speed on this hardware.
Ollama or llama.cpp (CUDA/ROCm) — vLLM if you need concurrent serving