{"ID":6267731,"CreatedAt":"2026-07-10T01:11:38.759438437Z","UpdatedAt":"2026-07-11T18:04:07.632774009Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.07836","arxiv_id":"2607.07836","title":"Infinity-Parser2 Technical Report","abstract":"We present Infinity-Parser2, a large multimodal model that couples a controllable data-synthesis pipeline with multi-task reinforcement learning for end-to-end document parsing, addressing the persistent scarcity of faithfully annotated parsing corpora. Our contributions are threefold. First, we build a scalable synthesis engine, pairing a controllable rendering framework with an iterative refinement loop, and use it to construct and open-source Infinity-Doc2-5M: a 5-million-sample bilingual (Chinese/English) corpus spanning diverse document types, annotated with element bounding boxes, canonical content forms (Markdown, HTML, LaTeX, SMILES, structured charts), and full-page reading order. Second, we introduce a verifiable, multi-task reward system that enables Joint Reinforcement Learning across eight co-trained objectives (document parsing, layout analysis, table parsing, math formula parsing, chart parsing, chemical formula parsing, document VQA, and general multimodal understanding), unifying perception, structure, and reasoning in a single optimization signal. Third, we release two variants under a shared architecture: Infinity-Parser2-Flash, optimized for low-latency inference with a $3.68\\times$ throughput gain over Infinity-Parser-7B, and Infinity-Parser2-Pro, engineered for precision-critical settings. Infinity-Parser2-Pro reaches state-of-the-art 87.6% on olmOCR-Bench and 74.3% on ParseBench, surpassing DeepSeek-OCR-2, PaddleOCR-VL-1.5, and MinerU2.5, with strong generalization to charts, chemical formulas, and document VQA.","short_abstract":"We present Infinity-Parser2, a large multimodal model that couples a controllable data-synthesis pipeline with multi-task reinforcement learning for end-to-end document parsing, addressing the persistent scarcity of faithfully annotated parsing corpora. Our contributions are threefold. First, we build a scalable synthe...","url_abs":"https://arxiv.org/abs/2607.07836","url_pdf":"https://arxiv.org/pdf/2607.07836v1","authors":"[\"Zuming Huang\",\"Jun Huang\",\"Kexuan Ren\",\"Baode Wang\",\"Weizhen Li\",\"Jianming Feng\",\"Yu Wang\",\"Yichen Yao\",\"Shijun Lin\",\"Yige Tang\",\"Cheng Peng\",\"Weidi Xu\",\"Wei Chu\",\"Yinghui Xu\",\"Yuan Qi\"]","published":"2026-07-08T18:17:21Z","proceeding":"cs.AI","tasks":"[\"cs.AI\"]","methods":"[\"Reinforcement Learning\"]","has_code":false}