Mistral Small

Mistral Small is a lightweight model developed by Mistral, positioned as the company's efficient option for applications requiring fast inference with reasonable capability. Within Mistral's model family, it serves as the speed-optimized tier below their more capable flagship offerings. The model supports text-based chat completion with a 32,000 token context window and includes tool calling functionality. It demonstrates strong inference speed with 123.31 output tokens per second and a 337ms time to first token, making it well-suited for applications where response latency matters. The model has a knowledge cutoff of September 2023. Mistral Small targets use cases where developers need reliable language model capabilities without the computational overhead of larger models. It competes with other lightweight offerings in the market by balancing performance with efficiency, making it suitable for production deployments requiring consistent speed.