Announcing MCP Toolbox Java SDK v1.0: Agentic data access for the enterprise
cloud.google.com·6d ago
Cloud & Infrastructure
The New Stack·Sep 1·5 min read
Meta just beat OpenAI and Google at real-time transcription
TL;DR
Meta's Superintelligence Labs has launched Muse Voice Transcribe, a cutting-edge real-time speech recognition model that outperforms competitors like OpenAI and Google in transcription accuracy.
✦ Why It Matters
Explore Muse Voice Transcribe for your transcription needs to enhance accuracy in your applications today.
Key Takeaways
- Consider using Muse Voice Transcribe for real-time transcription needs.
- Evaluate the model's performance against existing solutions in your projects.
- Leverage the API for multilingual transcription capabilities.
- Stay updated on competitive advancements in real-time transcription technology.
Full Summary
Muse Voice Transcribe, developed by Meta's Superintelligence Labs, is a cutting-edge real-time speech recognition model that excels in various benchmarks. It achieves a word error rate of 3.1% on the AA-WER Streaming benchmark, outperforming models from Cartesia, ElevenLabs, and OpenAI.
The model can recognize over 20 speakers and supports more than 70 languages, including multilingual conversations. It processes audio in 80-millisecond chunks, allowing it to adaptively manage latency based on word complexity.
This adaptive delay mechanism enhances transcription accuracy by providing context for difficult words. Muse Voice Transcribe is now available through the Meta Model API, priced at $3.00 per 1,000 audio minutes, but its weights will not be open-sourced.