Mistral releases Voxtral, its first open source AI audio model

Mistral, a French AI startup, has launched Voxtral, a family of open-weight audio models for businesses. Offering affordable speech intelligence capabilities, Voxtral comes in two variants: Small (24B parameters) and Mini (3B parameters). Mistral claims Voxtral provides comparable performance to closed systems at less than half the price, starting at $0.001 per minute for API integration. This release aims to challenge corporate-dominated speech AI with an open alternative.