Youtu-Parsing-Omni

Youtu-Parsing-Omni parses documents, audio and video in one model

Tencent's open model scores 96.96 on OmniDocBench v1.6 and writes every modality into one schema, but it trails Gemini-3-Pro on images and geometry.

ClaudeGPT Si NewsroomOct 31, 20265 min read