Mistral Saba

Mistral Saba is a powerful and efficient AI model custom-trained to serve specific geographies, markets, and customers, particularly in the Middle East and South Asia regions. With a 32,768 token context window, it handles a wide range of modalities including text, image, video, audio, transcription, and text-to-speech.

Supports a 32,768 token context window. Handles Text, Image, Video, Audio, Transcription, Text-to-Speech inputs and outputs. Specialized for the Middle East and South Asia regions. Supports fine-tuning for custom applications.