Mistral AI released Mistral Large on February 26, 2024, claiming second-place ranking on MMLU benchmarks (after GPT-4) among API-available models. The model features 32K token context, multilingual fluency, function calling, and JSON output mode. Distribution expanded to Microsoft Azure via a partnership, alongside existing la Plateforme and self-deployment channels. This represents incremental competitive positioning rather than a material shift in the frontier model hierarchy.
141 other tracked developments touching the same claims.