[key points]

Key points

• •Mistral AI has launched Voxtral Mini Transcribe V2 and Voxtral Realtime, two new speech-to-text models capable of translating between 13 languages.
• •Voxtral Realtime, designed for near real-time transcription within 200 milliseconds, is available under an open-source license and can run locally on devices.
• •Mistral aims to provide a cost-effective and less error-prone alternative to larger, general-purpose AI models, focusing on specialized models for specific tasks.
• •Mistral is positioning itself as a European-native, multilingual, open-source alternative to US-developed proprietary models, emphasizing compliance with EU regulations.