VibeVoice-ASR-Streaming, a unified streaming ASR model that continuously transcribes ''who said what'' as speech arrives,</p>\n","updatedAt":"2026-09-03T05:32:54.312Z","author":{"_id":"646c408336505117e22f4b36","avatarUrl":"/avatars/8201da3c4ec51c8e113446f9578d44f6.svg","fullname":"zhiliang","name":"zzliang","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":9,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.9596290588378906},"editors":["zzliang"],"editorAvatarUrls":["/avatars/8201da3c4ec51c8e113446f9578d44f6.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2609.02812","authors":[{"_id":"6a990641fea818274321ff9f","name":"Yujie Tu","hidden":false},{"_id":"6a990641fea818274321ffa0","name":"Zhiliang Peng","hidden":false},{"_id":"6a990641fea818274321ffa1","name":"Jianwei Yu","hidden":false},{"_id":"6a990641fea818274321ffa2","name":"Li Dong","hidden":false},{"_id":"6a990641fea818274321ffa3","name":"Songchen Xu","hidden":false},{"_id":"6a990641fea818274321ffa4","name":"Yaoyao Chang","hidden":false},{"_id":"6a990641fea818274321ffa5","name":"Wenhui Wang","hidden":false},{"_id":"6a990641fea818274321ffa6","name":"Zilong Wang","hidden":false},{"_id":"6a990641fea818274321ffa7","name":"Zehua Wang","hidden":false},{"_id":"6a990641fea818274321ffa8","name":"Yan Xia","hidden":false},{"_id":"6a990641fea818274321ffa9","name":"Jiajun Zhang","hidden":false},{"_id":"6a990641fea818274321ffaa","name":"Xie Chen","hidden":false},{"_id":"6a990641fea818274321ffab","name":"Furu Wei","hidden":false}],"publishedAt":"2026-09-02T00:00:00.000Z","submittedOnDailyAt":"2026-09-03T00:00:00.000Z","title":"VibeVoice-ASR-Streaming Technical Report","submittedOnDailyBy":{"_id":"646c408336505117e22f4b36","avatarUrl":"/avatars/8201da3c4ec51c8e113446f9578d44f6.svg","isPro":false,"fullname":"zhiliang","user":"zzliang","type":"user","name":"zzliang"},"summary":"Traditional speaker-attributed ASR systems treated ASR and speaker diarization as two separate tasks. Recently, end-to-end models such as VibeVoice-ASR have unified the two tasks within a single model. However, existing unified models still mainly support offline recognition, making it difficult to meet the low-latency requirements of real-time voice assistants and agents. To tackle this issue, we present VibeVoice-ASR-Streaming, one of the first LLM-based end-to-end approaches to streaming speaker-attributed ASR. It interleaves fixed-size audio chunks, a small amount of lookahead audio and previous text. This allows the model to produce ''who said what'' as speech arrives, without a separate diarization stage. For transcription accuracy, our 7B model achieves the lowest average WER/CER across five evaluation sets. For speaker attribution, it achieves the best or tied-best on 12 of 13 evaluation settings. We release the 1.5B and 7B model weights together with inference code.","upvotes":8,"discussionId":"6a990641fea818274321ffac","ai_summary":"A streaming, LLM-based end-to-end model unifies speaker-attributed speech recognition and diarization for low-latency real-time applications.","ai_keywords":["speaker-attributed ASR","end-to-end","LLM-based","streaming","speaker diarization","lookahead audio","WER/CER"],"ai_summary_model":"thinkingmachines/Inkling-Small","organization":{"_id":"5e6485f787403103f9f1055e","name":"microsoft","fullname":"Microsoft","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/1583646260758-5e64858c87403103f9f1055d.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"646c408336505117e22f4b36","avatarUrl":"/avatars/8201da3c4ec51c8e113446f9578d44f6.svg","isPro":false,"fullname":"zhiliang","user":"zzliang","type":"user"},{"_id":"68ac9da42166e79cec1b1571","avatarUrl":"/avatars/e2e138a8df6026e6f26df54ec32ed47a.svg","isPro":false,"fullname":"hulu","user":"hululuhu","type":"user"},{"_id":"6a97f937032af4b22d34bb54","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/xsK8NqdfHCu_IGRjdsb4z.jpeg","isPro":false,"fullname":"Zhiliang Peng","user":"zhiliangp","type":"user"},{"_id":"67ecd6178647cfa1775f75ed","avatarUrl":"/avatars/98882cc58dc0a5de94df765d523d92c9.svg","isPro":false,"fullname":"Furu Wei","user":"frontierai","type":"user"},{"_id":"69ccb304d5dcac9a4ac1772d","avatarUrl":"/avatars/826ea095e4b0eef72e147e10d1d06c39.svg","isPro":false,"fullname":"Isabella Wright","user":"victoriapqj64","type":"user"},{"_id":"63ac5701c21e60a3e9b58aa7","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63ac5701c21e60a3e9b58aa7/g6EX7diOpuA94R2ab-rZC.png","isPro":true,"fullname":"Dipankar Sarkar","user":"dipankarsarkar","type":"user"},{"_id":"67320e4996d5da4801a69199","avatarUrl":"/avatars/bf85472384886c19121a5b32bb4dbeea.svg","isPro":false,"fullname":"AlexTYJ","user":"AlexTYJ","type":"user"},{"_id":"620783f24e28382272337ba4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/620783f24e28382272337ba4/zkUveQPNiDfYjgGhuFErj.jpeg","isPro":false,"fullname":"GuoLiangTang","user":"Tommy930","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"5e6485f787403103f9f1055e","name":"microsoft","fullname":"Microsoft","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/1583646260758-5e64858c87403103f9f1055d.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2609/2609.02812.md","query":{}}">

Get this paper in your agent:

Don't have the latest CLI?

Cite arxiv.org/abs/2609.02812 in a dataset README.md to link it from this page.

Add this paper to a collection to link it from this page.