This browser is not keeping saved items, so they will be forgotten when you leave this page. Everything else still works.
Reference · Open-weight model family
Mistral
12 facts collected from their sources, each linked with its date. Nothing here is written from memory; a fact appears only as its source states it.
Models on Hugging Face
- mistralai/Timestral-123M is listed on Hugging Face (time-series-forecasting, timestral, license:other). Mistral AI on Hugging Face,
- mistralai/Timestral-28M is listed on Hugging Face (time-series-forecasting, timestral, license:other). Mistral AI on Hugging Face,
- mistralai/Timestral-5M is listed on Hugging Face (time-series-forecasting, timestral, license:other). Mistral AI on Hugging Face,
- mistralai/Timestral-209M is listed on Hugging Face (time-series-forecasting, timestral, license:other). Mistral AI on Hugging Face,
- mistralai/LIDstral-Arabic is listed on Hugging Face (text-classification, fasttext, license:apache-2.0). Mistral AI on Hugging Face,
- mistralai/Voxtral-Mini-4B-Realtime-Arabic is listed on Hugging Face (automatic-speech-recognition, vllm). Mistral AI on Hugging Face,
Prices
- mistral/codestral-latest at mistral: $0.3 input and $0.9 output per million tokens (USD). LiteLLM model price map,
Benchmark results
- mini-SWE-agent with devstral-2512: 53.8% (resolved) on SWE-bench, Verified. SWE-bench leaderboard,
- mini-SWE-agent with devstral-small-2512: 56.4% (resolved) on SWE-bench, Verified. SWE-bench leaderboard,
- SWE-agent with devstral-small-2507: 38% (resolved) on SWE-bench, Verified. SWE-bench leaderboard,
- OpenHands with mistralai/Devstral-Small-2505: 46.8% (resolved) on SWE-bench, Verified. SWE-bench leaderboard,
- Codestral 25.01: 11.1% (pass rate, second attempt) on Aider polyglot, 225 exercises. Aider polyglot leaderboard,