Youtu-Parsing-Omni

Youtu-Parsing-Omni is a compact parsing model with 5 billion parameters from Tencent's Youtu Lab. Its model card says a single input goes in and one structured JSON block comes out, following a schema the team calls OmniSchema, with coordinates on a 0 to 1000 grid and timestamps for audio and video. Accepted inputs are document pages, natural images, charts, flowcharts, geometry figures, audio clips and video, including video with a lot of text. The weights are published on Hugging Face and can be served with vLLM or with Transformers, the latter requiring remote code to be trusted. The licence is a custom one named Youtu-Parsing, and its opening lines state that the model is not intended for use within the European Union, a clause that prevails over the rest, so a European buyer should read it before any use. The card reports an overall score of 96.96 on OmniDocBench v1.6 and 75.08 on the team's own OmniParsingBench; these are Tencent's own figures and are not stored as scores. The technical report is listed as coming soon and the evaluation code as to be announced.

Not yet independently verified: The release date of 9 October 2026 rests only on the creation of the Hugging Face repository; the model card dates its report and weights only as October 2026 and lists the report as coming soon. The benchmark figures are Tencent's own, from the model card, so no score is stored. The licence is a custom one that states the model is not intended for use within the European Union. No independent report was found. We will update this when it can be confirmed, and remove this note.