Tencent open-sources WeVisDoc, tops OmniDocBench at 95.38

Tencent's WeVisDoc, an open-source end-to-end document parsing model (2B and 4B), converts a full page image into Markdown in one pass. Using a two-stage data-centric framework, it tops OmniDocBench with 95.38, with the 4B version leading mainly in formula and table parsing.

2026-09-18 ~ 2026-09-19 · 3 related posts

• Tencent's WeVisDoc tops OmniDocBench with 95.38 via two-stage data-centric training — tencent · 2026-09-18
• Tencent open-sources WeVisDoc: end-to-end document parsing model turns a page image into Markdown — xiaohu · 2026-09-19
• WeVisDoc-4B leads end-to-end document parsing, but its edge comes mostly from formulas and tables — xiaohu · 2026-09-19