Voxtral Mini 4B Realtime

This is the streaming-transcription member of the Voxtral family, architecturally distinct from the batch-oriented Voxtral Small 24B and Voxtral Mini Transcribe V2. Instead of requiring complete audio segments (padded to 30-second chunks inherited from the original Whisper encoderThe part of a model that reads input and converts it into an internal numerical representation the model can work with. In a translation model, the encoder reads the English sentence; the decoder produces the French. Modern chat models like GPT and Llama are "decoder-only" — they skip the separate encoder step. design), this model processes audio incrementally as it arrives. The practical use case is real-time voice agents — customer-service bots that respond mid-utterance, live captioning for video streams, and dictation applications where words appear as they're spoken. For batch transcription of complete recordings (podcasts, meetings), Voxtral Mini Transcribe V2 is cheaper and higher-quality. For true real-time streaming, this is the purpose-built option.