Trelis/tiron · Hugging Face
Log In<br>Sign Up
","eos_token":"","pad_token":"","unk_token":""}},"createdAt":"2026-07-18T21:57:12.000Z","discussionsDisabled":false,"discussionsSorting":"recently-created","downloads":22,"downloadsAllTime":22,"id":"Trelis/tiron","isLikedByUser":false,"availableInferenceProviders":[],"showHuggingChatEntry":false,"inference":"","lastModified":"2026-07-22T11:39:48.000Z","likes":7,"pipeline_tag":"automatic-speech-recognition","library_name":"transformers","librariesOther":[],"trackDownloads":true,"model-index":null,"private":false,"repoType":"model","gated":false,"tags":["transformers","safetensors","whisper","automatic-speech-recognition","multi-speaker","speaker-diarization","meeting-transcription","asr","en","zh","multilingual","license:apache-2.0","endpoints_compatible","region:us"],"tag_objs":[{"id":"automatic-speech-recognition","label":"Automatic Speech Recognition","type":"pipeline_tag","subType":"audio"},{"id":"transformers","label":"Transformers","type":"library"},{"id":"safetensors","label":"Safetensors","type":"library"},{"id":"en","label":"English","type":"language"},{"id":"zh","label":"Chinese","type":"language"},{"id":"multilingual","label":"multilingual","type":"language"},{"id":"whisper","label":"whisper","type":"other","clickable":true},{"id":"multi-speaker","label":"multi-speaker","type":"other","clickable":true},{"id":"speaker-diarization","label":"speaker-diarization","type":"other","clickable":true},{"id":"meeting-transcription","label":"meeting-transcription","type":"other","clickable":true},{"id":"asr","label":"asr","type":"other","clickable":true},{"id":"endpoints_compatible","label":"Inference Endpoints","type":"other","clickable":true},{"id":"license:apache-2.0","label":"apache-2.0","type":"license"},{"type":"region","label":"🇺🇸 Region: US","id":"region:us"}],"transformersInfo":{"auto_model":"AutoModelForSpeechSeq2Seq","pipeline_tag":"automatic-speech-recognition","processor":"AutoProcessor"},"safetensors":{"parameters":{"BF16":1543539200},"total":1543539200,"sharded":false,"totalFileSize":3087229512},"hasBlockedOids":false,"region":"us","isQuantized":false},"discussionsStats":{"closed":0,"open":1,"total":1},"query":{},"inferenceContextData":{"billableEntities":[],"entityName2Providers":{}},"hasQuantizations":false,"copyToBucketNamespaces":[]}">
Tiron
Released 21 July 2026.
Watch the launch video here. Try via API here.
Tiron is a state-of-the-art multi-speaker meeting transcription model. It jointly transcribes and attributes speech to speakers in a single decoding pass: for each 30-second audio window it emits an inline transcript with turn markers (up to 8 speakers per window) and timestamps.
Tiron uses the Whisper large-v3 architecture with an extended token vocabulary ( … , ). It is a drop-in WhisperForConditionalGeneration checkpoint.
Tiron is state of the art for whole-meeting transcription — among systems that reliably transcribe entire meetings, it leads every test set we evaluated (see Benchmarks).
For whole meetings (beyond a single 30s window), use the open-source harness at TrelisResearch/tiron, which adds chunking, cross-window speaker linking (ECAPA voice embeddings), and SRT/VTT/JSON output.
Benchmarks
Tiron is state of the art for whole-meeting transcription: among systems that can reliably transcribe an entire meeting end to end, it leads every test set we evaluated.
The closest such rival is AssemblyAI universal-3-pro (in its best diarized configuration), so we report the head-to-head against it below. Google's Gemini Pro is a strong short-clip transcriber — competitive with Tiron on short (~6 min) meetings — but degrades sharply on longer audio, truncating or failing outright on the longer ICSI and NOTSOFAR meetings, so it cannot be scored across full meeting corpora.
Pooled corpus cpWER (lower is better) on held-out meeting test sets, scored with identical references and normalization for every system:
Test set<br>AssemblyAI universal-3-pro<br>Tiron
AMI (4 meetings)<br>39.49<br>35.24<br>−4.3 pts (−11%)
ICSI (3 meetings)<br>34.64<br>20.91<br>−13.7 pts (−40%)
NOTSOFAR-1 (10 meetings)<br>38.62<br>37.55<br>−1.1 pts (−3%)
Macro (mean of corpora)<br>37.58<br>31.23<br>−6.4 pts (−17%)
Meetings evaluated (whole-meeting audio, far-field where applicable):
AMI : ES2004a, IS1009a, TS3003a, EN2002a
ICSI : Bmr013, Bmr018, Bro021
NOTSOFAR-1 : MTG_32040, MTG_32063, MTG_32072, MTG_32074, MTG_32092, MTG_32179, MTG_32185, MTG_32256, MTG_32257, MTG_32322
Scoring notes: cpWER is concatenated-permutation WER over whole meetings (transcription and speaker-attribution errors both count). Each corpus figure is pooled — total errors ÷ total reference words across that corpus's meetings — and the macro is the mean of the three corpus figures. On NOTSOFAR-1, stretches the human annotators marked (unintelligible) are masked from both hypothesis and reference for every system, so no system is rewarded for staying silent there. AMI and ICSI are unaffected by this mask.
Numbers are...