title: "Mistral batch API"

Mistral AI has introduced a new batch API designed for high-volume, asynchronous requests at a 50% lower cost than synchronous calls. This service is ideal for applications that prioritize data volume over immediate responses, such as bulk document processing or sentiment analysis. Users can upload a batch file of requests and download an output file once the processing is complete. The batch API is available for all models on the La Plateforme service and supports applications like document summarization, translation, and vector embedding.