Voxtral Small 24B 2507

Voxtral Small 24B 2507 is Mistral AI's 24-billion-parameter audio-language model built on Mistral Small 3, adding native audio input for speech transcription, translation, and audio understanding while preserving strong text generation capabilities. It is a general-purpose multimodal model that can listen to audio and produce text, summaries, or reasoned answers about audio content. Source: supabase.