AI BriefingSoMarkAI releases OXR-1.0, a 1.3B parameter open-source document parsing model
·2026.11.07 04:08
Key point
The 1.3B parameter model uses a recursive self-improvement loop and pluggable layout analysis to convert PDFs into structured data.
Details
- SoMarkAI released OXR-1.0, an open-source model that converts PDFs and document images into Agent-Friendly Markdown or JSON with layout coordinates.
- OXR ranks #1 on olmOCR-Bench, OmniDocBench, and DocParsingBench, with a specific lead in table parsing.
- The model features a pluggable layout analysis module, allowing users to replace the built-in layout model while retaining content recognition and reading order capabilities.
- OXR uses a recursive self-improvement (RSI) loop during post-training, a first in document parsing, and achieves top-tier speed on the RTX PRO 6000.
- The model recognizes 13 document element types, including formulas, tables, chemical structures, and code blocks, with bounding box coordinates.
- SoMarkAI provides an ecosystem of tools including SoMarkDown, SoPDF, and SDKs for Python and JS/TS.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Thinking Machines AI introduces Inkling 2 Small, planning Apache 2.0 open weights release next week