Mistral Launches Voxtral Speech Recognition Model
CHINA, JUL 14 – Kimi K2 offers advanced coding with a 66% success rate on benchmarks and is priced up to 90% cheaper than competitors, promoting open-source AI innovation worldwide.
- Mistral released Voxtral, its first family of open automatic speech recognition models aimed at businesses, on July 15, 2025.
- The release follows industry challenges with open-source speech models that trade off accuracy or semantic understanding, encouraging Mistral to bridge this gap.
- Voxtral offers multilingual transcription for up to 30 minutes of audio, supports queries and summaries, includes voice-activated function-calling, and comes in a 24B parameter production-scale and 3B parameter edge deployment versions.
- Mistral claims Voxtral outperforms OpenAI's Whisper large-v3, Gemini 2.5 Flash, and ElevenLabs Scribe with state-of-the-art accuracy and costs $0.001 per minute via API, less than half comparable solutions.
- Mistral is seeking partners to add features like speaker identification, offers private infrastructure deployment, aims for business adoption, and is in talks to raise up to $1 billion from investors such as Abu Dhabi's MGX fund.
20 Articles
20 Articles
: Apache-licensed plan takes aim at costlier options
Record number of unhoused people living in San Jose, new data shows
Alibaba-backed startup Moonshot released late Friday night its Kimi K2 model as a low-cost, open source large language model, with a focus on coding capabilities. Moonshot claimed Kimi K2 surpassed Claude Opus 4 on two benchmarks, and had better overall performance than OpenAI’s coding-focused GPT-4.1 model, according to several industry metrics. Coincidentally, OpenAI CEO Sam Altman announced early Saturday that there would be an indefinite del…
Mistral, a French start-up of artificial intelligence (IA), launched on Tuesday its first models centered on voice recognition and transcription in several languages. "The voice will be crucial in the future of human-machine interactions and will play a critical role in the adoption of digital technology.
Coverage Details
Bias Distribution
- 60% of the sources are Center
To view factuality data please Upgrade to Premium