# Meta Launches First Real-Time Audio Perception Model Muse Voice Transcribe

- Category: Flash (快讯)
- Published: 2026-09-02T11:52:00.000Z
- Source date: 2026-09-02
- HTML: https://www.88z.ai/en/news/flash-20260902-b2ffa05c705c947c
- Markdown: https://www.88z.ai/en/news/flash-20260902-b2ffa05c705c947c.md
- Other language: https://www.88z.ai/news/flash-20260902-b2ffa05c705c947c.md

Meta has released Muse Voice Transcribe, its first real-time audio perception model, integrating streaming speech recognition, speaker diarization, and endpoint detection. Supporting over 70 languages and multi-track transcription for more than 20 speakers, the model uses an adaptive latency mechanism to balance speed and accuracy, ranking first on the Artificial Analysis streaming speech-to-text leaderboard.

## At a glance

- Supports over 70 languages and multi-track transcription for 20+ speakers
- Ranks first on Artificial Analysis leaderboard via adaptive latency mechanism

## The update

Meta has officially launched Muse Voice Transcribe, its first real-time audio perception model. The system deeply integrates three core technologies: streaming speech recognition, speaker diarization, and endpoint detection, enabling efficient processing of complex audio scenarios. In practice, it supports recognition in over 70 languages and features multi-track transcription capabilities for more than 20 simultaneous speakers, significantly improving efficiency in multi-speaker environments. To ensure optimal user experience, the model incorporates an adaptive latency mechanism that balances high recognition accuracy with fast real-time response. Leveraging this technical advantage, Muse Voice Transcribe has topped the Artificial Analysis streaming speech-to-text leaderboard, securing the first position and demonstrating its leading status in the field of real-time audio processing.

## Sources

- [IT之家](https://www.ithome.com/0/997/218.htm) (2026-09-02)

This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.