M
Devstral 2 2512OPEN
Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding.
Prices update automatically — checked daily against provider list prices.
- $0.88 per 1M tokens, assuming 3 input for 1 output the convention used by Artificial Analysis
- cheaper than 20% of the 191 open-weight models tracked here
- 15 of 19 Mistral AI models tracked here, from cheapest
Benchmarks & performance
Independent benchmark scores for Devstral 2 2512, measured by Artificial Analysis. Higher is better.
Models at this level cost $0.96 per 1M tokens (median of 21) — this one costs $0.88.
Cheaper and better on this index: GLM 5.3 Flash · Gemini 3.7 Flash · Gemini 3.7 Flash (batch) and 63 more
About this model
Devstral 2 2512 is an open-weight AI model by Mistral AI. You can download and self-host it for free; the prices below are hosted-API list prices, tracked daily, for when you prefer convenience over self-hosting.
Frequently asked questions
What is Devstral 2 2512?
Devstral 2 2512 is an AI language model from Mistral AI. It is open-weight: you can download it and run it on your own hardware, for free. It scores 19.2 on the Artificial Analysis intelligence index.
Is Devstral 2 2512 free?
The weights are free and open — you can self-host Devstral 2 2512 and pay nothing per token. If you prefer a hosted API, list prices are $0.44 per million input tokens and $2.2 per million output tokens.
What is Devstral 2 2512 good at?
Independent benchmarks from Artificial Analysis give it GPQA 59.4%, MMLU-Pro 76.2%, Humanity's Last Exam 3.6%, Long Context Reasoning 31.3%, LiveCodeBench 44.8%, SciCode 33.1%, AIME 2025 36.7%, IFBench 38.1%, τ²-Bench 24.9%, τ-Bench Banking 10.5%, Terminal-Bench 30.3%, Terminal-Bench Hard 18.9%. It is particularly used for code generation.
How fast is Devstral 2 2512?
It generates about 43.1 tokens per second, with a median 0.75s delay before the first token. Measured independently by Artificial Analysis.
Can I self-host Devstral 2 2512?
Yes. Devstral 2 2512 has open weights, so you can download it and run it on your own GPU or server with tools like Ollama, vLLM or llama.cpp — with no per-token cost.