Microsoft Releases VibeVoice for Speech Transcription
The new open-source automatic speech recognition model handles multilingual transcription and speaker identification out of the box.
Microsoft has released VibeVoice-ASR, a new foundational model for automatic speech recognition. The system, now available on Hugging Face, is designed to convert spoken audio into written text across multiple languages.
Beyond simple transcription, VibeVoice's key capability is integrated speaker diarization—the ability to identify and label who is speaking and when. This feature is crucial for accurately transcribing conversations with multiple participants, such as meetings, interviews, or panel discussions, without requiring a separate post-processing step.
Why It Matters
The release adds a notable new entry into the competitive open-source audio space, which includes popular models like OpenAI's Whisper. While Microsoft has not yet published detailed performance benchmarks, VibeVoice’s built-in diarization offers a more streamlined solution for developers who would otherwise need to combine separate models for transcription and speaker identification.
Prospective users should take note of the licensing. According to the official model card, VibeVoice-ASR is being released for research purposes only. This will limit its immediate use in commercial products but provides a valuable new tool for the academic community exploring advanced speech processing systems.
Sources
- Visit
microsoft/VibeVoice-ASR
Hugging Face
More from Microsoft
All Microsoft releases →
Microsoft's Mage-VL-Cowork takes on computer-use tasks
A new vision-language agent model aims to operate GUIs, write code, and work the terminal.
Microsoft Releases FrogNano-4B Under MIT License
The compact 4-billion-parameter text model lands on Hugging Face with a permissive license aimed at developers.
Microsoft's VibeVoice ASR brings streaming speech-to-text
A 7-billion-parameter model targets real-time, multilingual transcription with an open release on Hugging Face.
More in Speech → Text
All Speech → Text →IBM's Granite Speech 5.0 Adds Speaker Attribution
The Apache-2.0 speech model handles transcription, translation, and figuring out who said what.
Qwen-Audio-3.1-ASR Adds Speaker-Aware Transcription
Alibaba's new open audio-language model transcribes and diarizes speech across 30 languages under an Apache 2.0 license.
NVIDIA's Nemotron 3.5 Transcribe Brings Promptable ASR
A sparse 30B mixture-of-experts model handles multilingual transcription, translation, and speaker diarization from a single system.
0 comments
No comments yet. Be the first to weigh in.