French artificial intelligence firm Mistral has introduced its latest model, Mistral Medium 3, designed to deliver exceptional efficiency while maintaining top-level performance. This new offering is now part of Mistral’s API with pricing set at forty cents per million input tokens and two dollars per million output tokens, a figure that competes favorably with pricier alternatives from industry rivals.
According to Mistral, Medium 3 consistently achieves results on AI benchmarks equal to or better than models like Anthropic’s Claude Sonnet 3 point 7, surpassing open peers such as Meta’s Llama 4 Maverick and Cohere’s Command A. Tokens play a vital role as data units for these models, with a million tokens representing an impressive seven hundred fifty thousand words.
Mistral Medium 3 Deployment and Reach
Flexibility defines Mistral Medium 3, which can operate on various cloud services or local infrastructure as long as it has access to at least four GPUs. The model’s affordability also outshines recognized cost leaders like DeepSeek v3, both in hosted APIs and self-managed deployments.
Launched in twenty twenty-three, Mistral has already established itself as a major force, developing AI-driven solutions such as its Le Chat chatbot and companion mobile apps. Since its founding, it has attracted significant investment, amassing over one point two billion dollars, and counts leading companies in sectors like finance and insurance among its clients.
The development team highlights Mistral Medium 3’s capabilities in coding, science, technology, engineering, and math, along with strong support for multimodal tasks. Enterprises from financial services to healthcare have already tested the model, applying it to activities including customer engagement, process automation, and complex data analysis.
Beyond its traditional API, enterprise AI users can deploy Mistral Medium 3 via collaborations with Mistral or on Amazon Sagemaker starting Wednesday, with accessibility soon expanding to Microsoft Azure AI Foundry and Google Vertex AI. This growing ecosystem reinforces the company’s strategic focus on making its technology widely available.
The recent release comes after Mistral Small 3 point 1 which debuted in March, and the company is teasing an even larger model in the near future. Further broadening its business products, Mistral has officially launched Le Chat Enterprise, a specialized chat platform featuring AI agent creation and native integration with various productivity tools.
Initially offered in private preview, Le Chat Enterprise becomes widely available today and will add support for the MCP standard, improving connectivity between AI assistants and operational platforms. Industry giants like Google and OpenAI have also committed to adopting MCP, aligning with Mistral’s move to accelerate enterprise adoption.
