Available Models
Mistral Small 3.2 24B 24B model optimized for instruction following and function calling. 128K context with strong coding, STEM, and vision capabilities. Excellent price-performance ratio.
Mistral Nemo Mistral’s lightweight 12B model co-developed with NVIDIA. 128K context at lowest cost. Ideal for simple tasks and high-volume applications.
Mistral Small 3.1 24B 24B parameter model with 128K context. Balanced performance for coding, instruction following, and general tasks. Apache 2.0 licensed.
Devstral 2512 Mistral’s specialized coding model optimized for software development tasks. Excellent for code generation, debugging, and technical assistance.
Devstral 2512 (Free) Free tier of Mistral’s Devstral coding model. Same capabilities with rate limits. Perfect for trying out coding assistance.