Autor: Jakub Rusinowski · Ostatnia aktualizacja: 15 sierpnia 2026
Model library → Devstral → Devstral Small 2505 24B
The first-generation Devstral Small (release 2505), and still the local coding model with the clearest published number attached to it: 46.8% on SWE-Bench Verified, which beat the prior open-source state of the art by 6 points at release. Finetuned from Mistral Small 3.1, so it inherits the 128K context. At 24B dense it is light enough for a single RTX 4090 or a 32GB Mac (~14 GB at Q4_K_M). Apache 2.0.
Devstral Small 2505 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 | 128,000 |
| Architecture | Dense Transformer |
| Provider | Mistral AI |
| Licence | Apache 2.0 |
| Specified at | Q4_K_M |
| System RAM | 32 GB |
| Record updated | 2026-08-15 |
Apache-2.0 — commercial use permitted. Commercial use permitted. No usage restrictions beyond attribution.
Modelled on a reference NVIDIA RTX 4090 (24 GB). Assumes an 8K-token context with an f16 KV cache. A longer window needs more; a quantized KV cache needs less. 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 | 10 GB | ~74 tok/s (est.) | Fits comfortably |
| Q3_K_M | 3.41 | 10.2 GB | 12.4 GB | ~60 tok/s (est.) | Fits comfortably |
| Q4_K_M | 4.83 | 14.5 GB | 16.6 GB | ~45 tok/s (est.) | Fits comfortably |
| Q5_K_M | 5.67 | 17 GB | 19.2 GB | ~39 tok/s (est.) | Fits comfortably |
| Q6_K | 6.56 | 19.7 GB | 21.8 GB | ~34 tok/s (est.) | Tight fit |
| Q8_0 | 8.50 | 25.5 GB | 27.6 GB | ~4 tok/s (est.) | Offloads to system RAM (slow) |
| F16 | 16.00 | 48 GB | 50.1 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 2505 24B is the AMD Radeon RX 9060 XT 16GB (16 GB).
Install Ollama, then run:
ollama run devstral:24bWeights on Hugging Face: mistralai/Devstral-Small-2505.
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 | 46.8 / 100 % | reported · https://mistral.ai/news/devstral/ |
Best for: software engineering, agentic coding, consumer gpu, local inference.
← All Devstral models | VRAM calculator | Check your own hardware