Youtu-Parsing-Omni
Youtu-Parsing-Omni is a 5B omni-modal parsing model that accepts documents, images, charts, geometry figures, audio, and video. It produces structured JSON covering perception and cognition tasks, with outputs selected by a task prompt.
What it does
Use it to extract structured content such as text, tables, formulas, timestamps, speech transcripts, and visual descriptions from varied media.
- Availability
- Model weights, vLLM plugin and inference examples released; available on Hugging Face and ModelScope
- Pricing model
- Not established
- Deployment
- vLLM plugin and inference examples; CUDA GPUs for inference
- Privacy
- Not established
- Licence
- license: other
Capability
Modality
Access
Deployment
Pricing
Target user
Evidence
6 claims- The 5B model parses documents, natural images, charts, flowcharts, geometry figures, audio, and video into structured JSON. [1]
<p align="center"><b>One model that parses documents, images, charts, geometry, audio and video into structured JSON</b></p>
Tencent Hunyuan. Quoted text is outside CC BY 4.0 and keeps its original rights. - Its outputs can include layout elements, text, tables, formulas, bounding boxes, timestamps, ASR, acoustic events, and descriptions. [1]
formulas, bounding boxes, timestamps, ASR, OCR, acoustic events, camera motion)
Tencent Hunyuan. Quoted text is outside CC BY 4.0 and keeps its original rights. - Model weights, a vLLM plugin, and inference examples are released. [1]
- **[2026-10]** [Technical report](paper/Youtu_Parsing_Omni.pdf), model weights, vLLM plugin and inference examples released.
Tencent Hunyuan. Quoted text is outside CC BY 4.0 and keeps its original rights. - The model is available from Hugging Face and ModelScope. [1]
- The quick start specifies Python 3.10 or later and CUDA GPUs for inference. [1]
# choose one mode per Python environment (Python >= 3.10, CUDA GPUs for inference)
Tencent Hunyuan. Quoted text is outside CC BY 4.0 and keeps its original rights. - The source reports an OmniDocBench v1.6 overall score of 96.96 and an OmniParsingBench average of 75.08. [1]
- **Strong results at a small size** – state-of-the-art on OmniDocBench v1.6 (96.96 Overall), best open-weight model on OmniParsingBench (75.08 Avg., second only to Gemini-3-Pro), and competitive with specialized models on chemical-structure (ChemOCR) and music-score (PDMX-Synth) recognition ([results](#results)).
Tencent Hunyuan. Quoted text is outside CC BY 4.0 and keeps its original rights.
Sources
1
1
Tencent HunyuanOwner-controlled
https://huggingface.co/tencent/Youtu-Parsing-Omni
Retrieved 2026-10-28 00:39 UTC
Verified
Versions
3| Version | Effective | Change | Changed fields | State |
|---|---|---|---|---|
| 3 | 2026-10-11 13:49 UTC | material change | capability, availability, deployment | Current |
| 2 | 2026-10-10 14:45 UTC | material change | capability, availability, deployment | Current |
| 1 | 2026-10-09 15:42 UTC | publication | None | Current |
In briefs
2- Turn a prompt into an editable slide, poster, or dashboard Weekly brief 2026-W41
- Turn documents, charts, audio, and video into structured information Daily brief 2026-10-09
Original catalog content is licensed under CC BY 4.0, attributed to wowsuchai. Source quotations and linked content keep their original rights.