Post-Cutoff
Search For AI agents

MicrosoftVibeVoice familyCurrentSpeech and audioReleased 25 Aug 2025

VibeVoice (ASR, ASR-Streaming, ASR-BitNet, Realtime-0.5B TTS)

Open Microsoft speech models for long-form ASR and lightweight streaming TTS.

Model idmicrosoft/VibeVoice-ASR

PricePrice not published

Checked29 September 2026

Where it stands

89newer speech and audio models since its release

The 5 newest are listed. All speech and audio models

After it from Microsoft
MAI-Voice-2 / MAI-Voice-2-Flash2 Jun 2026
Weights
Weights are public under a permissive licence (MIT).

How to call it

7 ways to use VibeVoice (ASR, ASR-Streaming, ASR-BitNet, Realtime-0.5B TTS), with the model id each provider expects.

Providers, model ids, endpoints and docs for VibeVoice (ASR, ASR-Streaming, ASR-BitNet, Realtime-0.5B TTS)
ProviderModel idEndpointDocs
Hugging Face microsoft/VibeVoice-ASR huggingface.co
Hugging Face (Transformers format) microsoft/VibeVoice-ASR-HF huggingface.co
Hugging Face microsoft/VibeVoice-ASR-BitNet huggingface.co
Hugging Face microsoft/VibeVoice-Realtime-0.5B huggingface.co
Hugging Face (streaming ASR, 7B repo; 9B params incl. decoder) microsoft/VibeVoice-ASR-Streaming-7B huggingface.co
Hugging Face (streaming ASR, small) microsoft/VibeVoice-ASR-Streaming-1.5B huggingface.co
GitHub github.com

Pricing

Price not published.

No price source linked. Checked 29 September 2026.

Specs

Input
Audio and text
Output
Text and audio
Context window
Not published
Max output
Not published
Open weights
Yes
Licence
MIT
Released
25 August 2025
Status
Current
Type
Speech and audio

What stands out

What its maker marketed at launch, and what people found later, each with its source.

  • 60-minute single-pass ASR with diarization

    VibeVoice-ASR (~9B params incl. Qwen2-based decoder) transcribes up to 60 min in one pass with who/when/what structured output, hotwords and 50+ languages with code-switching.

    Source: huggingface.co

  • CPU-only realtime ASR

    Found after launch

    VibeVoice-ASR-BitNet (2026-07-23) compresses the model 4.62 GB -> 1.58 GB and runs faster than real time on 3 CPU threads (1.6-2.3x faster than Whisper.cpp).

    Source: huggingface.co

  • Streaming speaker-attributed ASR

    Found after launch

    VibeVoice-ASR-Streaming (7B and 1.5B repos, uploaded 2026-09-02) transcribes live audio with speaker attribution (who said what) and custom hotwords in 10 languages (zh, en, fr, de, it, ja, ko, pt, ru, es); MIT license.

    Source: huggingface.co

  • Long-form multi-speaker TTS (withdrawn)

    Found after launch

    Original VibeVoice-TTS (1.5B/7B) generated up to 90 min with 4 speakers; Microsoft removed the TTS code on 2025-09-05 over responsible-AI misuse concerns.

    Source: github.com

Notes

Open-source voice research family from Microsoft (MIT). Timeline: TTS 2025-08-25 (code pulled 2025-09-05), Realtime-0.5B streaming TTS (~300 ms first audio) 2025-12-03, ASR 2026-01-21, Transformers integration 2026-03, Foundry Labs 2026-03-12, ASR-BitNet 2026-07-23, ASR-Streaming (10 languages, hotwords, speaker attribution) announced 2026-09-03; HF repos microsoft/VibeVoice-ASR-Streaming-7B and -1.5B created 2026-09-02 (verified 2026-09-29). Monthly downloads to 2026-09-29: VibeVoice-ASR ~734k, VibeVoice-1.5B ~717k. Separate from Microsoft’s proprietary MAI-Voice/MAI-Transcribe.

See the Hugging Face Transformers docs (model_doc/vibevoice_asr) and the GitHub repo for inference code.

Sources: https://github.com/microsoft/VibeVoice · https://huggingface.co/microsoft/VibeVoice-ASR · https://huggingface.co/docs/transformers/model_doc/vibevoice_asr