Get ready for a game-changer in speech-to-text technology! Voxtral Transcribe 2 is here, and it's revolutionizing the way we transcribe audio. With two cutting-edge models, Voxtral Mini Transcribe V2 and Voxtral Realtime, we're pushing the boundaries of what's possible. But here's where it gets controversial: these models offer state-of-the-art transcription quality, speaker diarization, and ultra-low latency, all at a fraction of the cost!
The Future of Transcription is Here!
Voxtral Mini Transcribe V2 is a powerhouse, delivering top-notch transcription in 13 languages with speaker diarization, context biasing, and precise word-level timestamps. And for live applications, Voxtral Realtime takes the lead with configurable latency down to sub-200ms, making it perfect for voice agents and real-time apps.
Efficiency Meets Excellence
The efficiency of these models is truly remarkable. Voxtral Mini Transcribe V2 achieves industry-leading accuracy at an incredibly low word error rate, making it the most cost-effective option on the market. And with open weights under the Apache 2.0 license, Voxtral Realtime can be deployed on edge devices for privacy-focused applications.
Realtime: Unlocking New Possibilities
Voxtral Realtime is a game-changer for low-latency applications. Unlike traditional methods, it uses a novel streaming architecture to transcribe audio as it arrives, offering customizable delays down to sub-200ms. This opens up a whole new world of voice-first applications.
Multilingual Mastery
Both models excel in multilingual transcription. Voxtral Realtime supports 13 languages, including English, Chinese, Hindi, and more, with strong performance across the board. And with a 4B parameter footprint, it runs efficiently on edge devices, ensuring privacy and security.
Try It Out: Audio Playground
We've launched an audio playground in Mistral Studio, where you can test Voxtral Transcribe 2 instantly. Upload your audio files, toggle diarization, choose timestamp options, and see the magic happen!
Transforming Voice Applications
Voxtral is powering a wide range of voice-based applications and industries. From meeting intelligence and voice agents to contact center automation and media broadcasting, Voxtral is making an impact. With its accurate transcription, speaker diarization, and low latency, Voxtral is transforming the way we interact with voice technology.
Get Started Today
Voxtral Mini Transcribe V2 is available now via API at an unbeatable price of $0.003 per minute. Voxtral Realtime is also accessible via API at $0.006 per minute, and its model weights are open-source on Hugging Face.
Explore our documentation to learn more about Voxtral's capabilities and join us in building the future of speech AI. We're always looking for talented individuals to join our team and make a difference.
So, what do you think? Are you ready to embrace the power of Voxtral Transcribe 2? We'd love to hear your thoughts and experiences in the comments below!
