Model Releases

Tencent Hunyuan releases HyOCR-1.5: end-to-end OCR large model fully open-sourced, inference 6.37x faster

Original title:训推全开源:HyOCR-1.5 让端到端 OCR 大模型跑得更快、看得更准、功能更全

AI summary

Tencent Hunyuan released HyOCR-1.5, the first expert model in the end-to-end OCR large model field to fully open-source training, inference, and model weights. With only 1B parameters, it covers more than 8 text-centric tasks. It introduces the DFlash speculative decoding framework, achieving 6.37x acceleration under Transformers and 2.14x under vLLM, with end-to-end inference at 1.408s per page. It supports 4K resolution and a 128K context window, and extends low-resource OCR (331 languages), ancient script recognition, and multi-image question answering through Agentic Data Flow. It ranks first among end-to-end models on OmniDocBench v1.6 with a score of 94.74.

Key Takeaways

  • Tencent Hunyuan released HyOCR-1.
  • 5,the first expert model in the end-to-end OCR large model field to fully open-source…
  • With only 1B parameters,it covers more than 8 text-centric tasks.
  • It introduces the DFlash speculative decoding framework,achieving 6.
  • 37x acceleration under Transformers and 2.

Quick Facts

Published
Updated
Source
公众号:腾讯混元
Category
Model Releases
AI score
76
Read the original Read the full article at the original source
Why it matters

我觉得这是近期最实在的OCR开源,1B参数覆盖多任务、推理加速6倍,还支持llama.cpp本地部署,做文档解析的同学可以直接上车。

Related entities

Source & copyright

This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.

Original article · AIHOT entry

How to Cite

When citing this article, keep the title, date, and permalink, and credit Orange News as the source.

Orange News: Tencent Hunyuan releases HyOCR-1.5: end-to-end OCR large model fully open-sourced, inference 6.37x faster. July 13, 2026. https://new.nat6.net/en/news/cmrj4s89p05nhbilkljqel2d5/

View all →