Devstral Small 2
Mistral AI
🧠 Language Model
Lightweight code agents model. 24B parameters, Apache 2.0 licensed. Fast and efficient.
Specifications
Context Window131,072 tokens
Max Output8,192 tokens
Speed●●●●● Fast
ModalitiesInput: text · Output: text
FeaturesTools: Yes Streaming: Yes
Pricing
Included with plan
Capabilities
More from Mistral AI
Mistral Large 3
Open-weight flagship. 675B MoE (41B active), multimodal with vision. Apache 2.0 licensed.
Mistral Small
Fast and efficient model for simple tasks. Great balance of performance and cost.
Codestral
Specialized model for code generation, completion, and understanding.
Magistral Medium
Advanced reasoning model for complex multi-step logic and problem-solving.