{"slug": "meta-superintelligence-labs-releases-muse-voice-transcribe-one-real-time-model", "title": "Meta Superintelligence Labs Releases Muse Voice Transcribe: One Real-Time Model for Streaming ASR, Diarization, and Endpointing", "summary": "Meta Superintelligence Labs released Muse Voice Transcribe, a real-time audio perception model that performs streaming ASR, speaker diarization for 20+ speakers, and endpointing in a single autoregressive pass, available as a hosted API at $3.00 per 1,000 audio minutes. The model, which powers dictation in Meta AI for Mac and Muse Code, achieves 3.1% final-transcript WER on the Artificial Analysis AA-WER Streaming benchmark and supports 70+ languages with 25 extensively verified at launch.", "body_md": "Most production voice stacks are three systems stitched together. One model transcribes, a second separates speakers, and a detector decides when the user stopped talking. Each hand-off adds latency and a new failure mode.\n\n[Muse Voice Transcribe](https://research.meta.ai/blog/introducing-muse-voice-transcribe), announced by Meta Superintelligence Labs this week, collapses those three jobs into a single autoregressive model. Meta calls it its first real-time audio perception model. It performs streaming ASR, speaker diarization for 20+ speakers, and endpointing in one pass, with no required post-processing.\n\n**Is it deployable?** Yes, but only as a hosted API. It is live on the [Meta Model API](https://developer.meta.com/ai/models/muse-voice-transcribe) as `muse-voice-transcribe-1.0`\n\nat $3.00 per 1,000 audio minutes ($0.18 per hour), and it already powers dictation in Meta AI for Mac and [Muse Code](https://developer.meta.com/ai/products/muse-code). No weights have been released, so there is no self-hosted path.\n\n**Streaming ASR as the foundation**\n\nMuse Voice Transcribe is an autoregressive multimodal model from the Muse Spark family. Audio arrives in 80ms chunks at 12.5 Hz. Each chunk is transformed into a single soft token.\n\nAfter every chunk the model makes one binary choice. It either predicts a `<|next_audio|>`\n\ntoken and keeps listening, or it emits a text token. When the model predicts `<|next_audio|>`\n\n, that token is replaced by the actual next audio chunk in the input. When the stream ends, an `<|empty_audio|>`\n\ntoken is inserted, and the model flushes all remaining text without requesting more audio.\n\nListening and writing share one decoder loop, so there is no separate alignment stage to drift.\n\n**Adaptive delay, trained with RL**\n\nBecause the model controls when it listens, it also controls how much audio context sits behind each word. Meta calls that gap ‘delay.’ Longer delay means a more accurate transcript and higher latency.\n\nInstead of fixing that trade-off, Meta trains it. Reinforcement learning combines a word error rate reward and a delay reward multiplicatively, producing a policy that varies delay per word by difficulty. Meta reports this puts the model on the Pareto front for speed against accuracy, measured by time to final transcription, ahead of the previous frontier formed by Soniox, Cartesia, and ElevenLabs systems.\n\n**Diarization and endpointing are more tokens**\n\nMeta did not add a second model for speaker attribution. It added special tokens to the same stream.\n\nFor diarization, a `<|start_of_turn|>`\n\ntoken marks a potential speaker switch, and a `<|speaker_{A-Z}|>`\n\ntag identifies the speaker. The turn token fires as soon as a switch is possible, while the speaker tag is delayed to the end of the chunk. Audio from one speaker can be split across several segments that all resolve to the same tag.\n\nFor endpointing, `<|speech_onset|>`\n\nmarks the start of speech and `<|speech_endpoint|>`\n\nmarks the point where the user finished. Both tasks are trained jointly with streaming ASR, using extra rewards layered on top of the ASR reward.\n\n**Capabilities**\n\nThe model was trained on 70+ languages, of which 25 are extensively verified and recommended at launch. Code-switching is native, both within a sentence and between sentences, which matters for bilingual speakers who mix languages mid-clause. Accuracy can be improved further with language, keyword, and context biasing.\n\nLong-context handling is a practical differentiator. Meta states the model natively supports audio input exceeding one hour and 20+ speakers, with no required post-processing step.\n\n**Benchmarks**\n\nMeta reports first place on Artificial Analysis for streaming speech-to-text and on public diarization benchmarks, as of September 1, 2026.\n\nOn [Artificial Analysis AA-WER Streaming](https://x.com/ArtificialAnlys/status/2094849283120128135), Muse Voice Transcribe records 3.1% final-transcript WER at 0.16s after end of speech. Cartesia Ink-2 with semantic endpoints is 3.4% at 0.43s. ElevenLabs Scribe v2 Realtime is 3.6% at 0.14s. Cartesia Ink-2 with external endpoints is fastest at 0.07s but least accurate at 4.0%. On first partial transcript, Muse Voice Transcribe records 3.6% WER at 0.13s.\n\nOn diarization, Meta reports a 17.5% average diarization error rate across AMI-IHM, AMI-SDM, and VoxConverse. Five other systems in the same chart range from 21.1% to 28.6%.\n\nPrice is the other axis. At $3.00 per 1,000 minutes, it undercuts Cartesia Ink-2 at $4.00 and is less than half the $6.50 for ElevenLabs Scribe v2 Realtime and Deepgram Flux.\n\n**Interactive explainer**\n\n**Key Takeaways**\n\n- Single model does streaming ASR, diarization for 20+ speakers, and endpointing.\n- 3.1% final-transcript WER at 0.16s on Artificial Analysis AA-WER Streaming.\n- Reinforcement learning teaches the model a per-word \"adaptive delay\" policy.\n- Trained on 70+ languages, 25 extensively verified, with native code-switching.\n- API-only at $0.18 per audio hour. No open weights.\n\n**Check out** [ Meta AI Research blog](https://research.meta.ai/blog/introducing-muse-voice-transcribe),\n\n[,](https://x.com/AIatMeta/status/2094839236016976028)\n\n**AI at Meta on X**[and](https://x.com/ArtificialAnlys/status/2094849283120128135)\n\n**Artificial Analysis**[. Also, feel free to follow us on](https://developer.meta.com/ai/models/muse-voice-transcribe)\n\n**Meta Model API model page****and don’t forget to join our**[Twitter](https://x.com/intent/follow?screen_name=marktechpost)\n\n**and Subscribe to**\n\n[150k+ML SubReddit](https://www.reddit.com/r/machinelearningnews/)**. Wait! are you on telegram?**\n\n[our Newsletter](https://magic.beehiiv.com/v1/f5e63dd4-5653-4f09-83e2-321a8b1ba526?email={{email}})\n\n[now you can join us on telegram as well.](https://t.me/machinelearningresearchnews)Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? [Connect with us](https://forms.gle/wbash1wF6efRj8G58)\n\nMichal Sutter is a data science professional with a Master of Science in Data Science from the University of Padova. With a solid foundation in statistical analysis, machine learning, and data engineering, Michal excels at transforming complex datasets into actionable insights.", "url": "https://wpnews.pro/news/meta-superintelligence-labs-releases-muse-voice-transcribe-one-real-time-model", "canonical_source": "https://www.marktechpost.com/2026/09/01/meta-superintelligence-labs-releases-muse-voice-transcribe-one-real-time-model-for-streaming-asr-diarization-and-endpointing/", "published_at": "2026-09-02 05:37:12+00:00", "updated_at": "2026-09-02 05:52:47.496782+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "natural-language-processing", "ai-products", "ai-research"], "entities": ["Meta Superintelligence Labs", "Muse Voice Transcribe", "Meta Model API", "Meta AI for Mac", "Muse Code", "Artificial Analysis", "Cartesia", "ElevenLabs"], "alternates": {"html": "https://wpnews.pro/news/meta-superintelligence-labs-releases-muse-voice-transcribe-one-real-time-model", "markdown": "https://wpnews.pro/news/meta-superintelligence-labs-releases-muse-voice-transcribe-one-real-time-model.md", "text": "https://wpnews.pro/news/meta-superintelligence-labs-releases-muse-voice-transcribe-one-real-time-model.txt", "jsonld": "https://wpnews.pro/news/meta-superintelligence-labs-releases-muse-voice-transcribe-one-real-time-model.jsonld"}}