French AI startup Mistral recently released its first open-source audio AI models, Voxtral, directly competing with proprietary systems from Google and OpenAI. The Paris-based company claims that Voxtral offers state-of-the-art speech understanding at less than half the price of existing APIs, aiming to democratize access to advanced voice intelligence. Released under an Apache 2.0 license, Voxtral includes multiple models tailored to different needs, such as the flagship Voxtral Small and the compact Voxtral Mini. Both are available for download on Hugging Face and can be integrated via a simple API call, starting at just $0.001 per minute. Mistral emphasizes accessibility with features like multilingual support and automatic language detection. The launch positions Voxtral not just as a technical innovation but also as a strategic response to the fierce competition in the AI voice market, promising to empower developers while maintaining affordability and performance.
Source link
Share
Read more