The best model in its weight class
- We’re committed to empower the AI community with open technology. Our open models sets the bar for efficiency, and are available for free, with fully permissive license. - Our optimized commercial models are designed for performance and are available via our flexible deployment options.
Mistral Small 3.1 is a lightweight, multimodal model optimized for cost-sensitive projects. At $0.1/M input tokens and $0.3/M output tokens, it delivers state-of-the-art performance for tasks like text generation, classification, and summarization while minimizing API costs compared to larger models.
Mistral Small 3.1 is available via Mistral's API and can be deployed through Studio, Vibe, or self-hosted infrastructure. Developers can integrate it into workflows with full observability, evals, and guardrails through Mistral Studio's unified registry and deployment portability.
Mistral Small 3.1 is released as open-weight under Apache 2.0 license for research/individual use, enabling self-hosting on any infrastructure. It supports multiple languages and modalities, allowing organizations to maintain control over their AI deployments and data.
Mistral Small 3.1 powers Vibe agents and can be orchestrated through Mistral Studio for multi-step task automation, persistent memory, and tool integration. It enables developers to build autonomous agents for document analysis, research, and complex reasoning workflows.