Mistral AI has introduced Voxtral, an open-source family of speech understanding models that go beyond simple transcription. Available in 24B and 3B sizes, Voxtral offers advanced features like question answering, summarization, and function calling directly from voice input, all with state-of-the-art performance. It supports multiple languages, including English, Spanish, French, Hindi, German, Dutch, and Italian, and can process audio up to 30 minutes long. Voxtral is available under the Apache 2.0 license, allowing for both commercial and private use. ([digit.in](https://www.digit.in/features/general/what-is-voxtral-mistrals-open-source-ai-audio-model-key-features-explained.html?utm_source=openai))