Autor: Jakub Rusinowski · Ostatnia aktualizacja: 11 września 2026
Model library → Devstral → Devstral Small 2 24B
The second-generation Devstral Small, and the reason this family matters to anyone with one GPU: 68.0% on SWE-bench Verified — up 21 points on the 2505 above — at the same 24B dense, about 15.3 GB at Q4_K_M on a single RTX 4090 or a 32 GB Mac. 256K maximum context and multimodal input, so "read this screenshot and fix the CSS" works. Apache 2.0, with a first-party Ollama tag.
Devstral Small 2 24B needs about 15 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache. On Apple Silicon that figure comes out of unified memory.
| Parameters | 24 Billion |
| Context window | 262,144 |
| Architecture | Dense Transformer (vision) |
| Provider | Mistral AI |
| Licence | Apache 2.0 |
| Specified at | Q4_K_M |
| System RAM | 32 GB |
| Record updated | 2026-09-11 |
Apache-2.0 — commercial use permitted. Commercial use permitted. No usage restrictions beyond attribution.
Modelled on a reference NVIDIA RTX 4090 (24 GB). Weights plus framework overhead only — this model publishes no architecture we can read, so no KV cache is included. A real session needs more; the figure is a floor, not a target. Speed figures are ESTIMATES from the memory-bandwidth roofline described on the methodology page, not benchmarks we ran — rows marked measured come from published or reader-submitted runs. VRAM here includes the KV cache, so it reads higher than the headline figure above, which does not.
| Quant | Bits/weight | Weights | VRAM needed | Est. speed | Fit on 24 GB |
|---|---|---|---|---|---|
| Q2_K | 2.63 | 7.9 GB | 8.7 GB | ~73 tok/s (est.) | Fits comfortably |
| Q3_K_M | 3.41 | 10.2 GB | 11 GB | ~59 tok/s (est.) | Fits comfortably |
| Q4_K_M | 4.83 | 14.5 GB | 15.3 GB | ~44 tok/s (est.) | Fits comfortably |
| Q5_K_M | 5.67 | 17 GB | 17.8 GB | ~39 tok/s (est.) | Fits comfortably |
| Q6_K | 6.56 | 19.7 GB | 20.5 GB | ~34 tok/s (est.) | Fits comfortably |
| Q8_0 | 8.50 | 25.5 GB | 26.3 GB | ~4 tok/s (est.) | Offloads to system RAM (slow) |
| F16 | 16.00 | 48 GB | 48.8 GB | ~2 tok/s (est.) | Offloads to system RAM (slow) |
Want to set your own context length and KV-cache quantization? Use the interactive VRAM calculator.
or compare on Vast.ai from $0.35/hr (typical low · varies)
As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.
The cheapest catalogued GPU that runs Devstral Small 2 24B is the AMD Radeon RX 9060 XT 16GB (16 GB).
Install Ollama, then run:
ollama run devstral-small-2:24bWeights on Hugging Face: mistralai/Devstral-Small-2-24B-Instruct-2512.
Quality scores as published by the model's authors or an independent evaluator — not throughput, and not measured by us.
| Benchmark | Score | Provenance |
|---|---|---|
| SWE-bench Verified | 68 / 100 % | reported · https://mistral.ai/news/devstral-2-vibe-cli/ |
Best for: software engineering, agentic coding, consumer gpu, multimodal.
← All Devstral models | VRAM calculator | Check your own hardware