Mistral: Mistral Medium 3.5

Mistral Medium 3.5 is a multimodal model from Mistral AI that accepts text, images, and files as input, with a 262,144-token context window. It supports tool use and reasoning, making it applicable to agentic workflows and multi-step tasks. Structured output support is unconfirmed from available data. At $1.50 per million input tokens and $7.50 per million output tokens, it sits in a mid-tier price range. Its blended benchmark score of 63.3 covers only 3 benchmarks, so treat that figure as preliminary rather than comprehensive. The standout data point is its agentic benchmark score of 87.7, which makes it worth shortlisting for teams building tool-calling or orchestration pipelines on a moderate budget. Teams needing strong general or coding performance should weigh the limited benchmark coverage carefully before committing.