document parsing
3 published articles
Analysis
The 0.8B model that just broke document parsing's pipeline monopoly
OvisOCR2 (0.8B parameters) from Tencent scored 96.58 on OmniDocBench, the first end-to-end model to top the leaderboard. The model uses reinforcement learning and distillation to outperform larger pipeline systems, challenging the necessity of multi-stage architectures.
2026-07-26
Document AI
The 0.8B model that beat every pipeline document parser at its own game
Alibaba's OvisOCR2, a compact 0.8B end-to-end model, achieves state-of-the-art scores on OmniDocBench (96.58) and PureDocBench (75.06), overtaking larger pipeline-based parsers. Its success comes from a data engine mixing filtered real documents with synthetic pages, reinforcement learning on a 4B teacher, and on-policy distillation into the small model.
2026-07-24
Document AI
Mistral OCR 4 knows where each word lives and how much to trust it
Mistral OCR 4 introduces structured document parsing with bounding boxes, block classification, and confidence scores. It beats leading OCR systems in human evaluations and benchmarks, supports 170 languages, and runs in a single container for self-hosted deployments.
2026-07-18