HunyuanOCR-1.5: Making Lightweight OCR VLMs Faster and Better

2026-07-06

AI Summary

This work addresses the limitations of existing lightweight end-to-end OCR vision-language models, which suffer from slow inference on long structured outputs—such as tables and mathematical expressions—and underperform on long-tail tasks like ancient script recognition and fine-grained diagram parsing. Building upon the HunyuanOCR-1.0 architecture, we propose DFlash to accelerate OCR decoding while preserving output distribution, achieving substantial latency reduction. We further introduce the first Agentic Data Flow mechanism, which autonomously converts model weaknesses into executable data requirements, enabling closed-loop self-directed processes for data retrieval, verification, and pipeline construction. Combined with high-resolution, long-context pretraining and post-training strategies, our system attains state-of-the-art performance on OmniDocBench v1.6, delivering 6.37× faster inference than standard Transformers (2.14× under vLLM), establishing the current fastest lightweight OCR VLM and achieving breakthrough results across multiple long-tail tasks.

Abstract

We present HunyuanOCR-1.5, a lightweight end-to-end OCR-specialized vision-language model. HunyuanOCR unifies document parsing, text spotting, information extraction, text-image translation, and multi-image document understanding within a single end-to-end VLM. Building upon the lightweight architecture of HunyuanOCR-1.0, HunyuanOCR-1.5 does not redesign the backbone, but systematically improves both efficiency and capability. For efficiency, we adapt DFlash to OCR decoding, significantly reducing the latency of long structured outputs such as dense documents, tables, and formulas while preserving output distribution. Powered by DFlash, HunyuanOCR-1.5 achieves a 6.37x Transformer inference speedup and a 2.14x speedup under vLLM, delivering the fastest inference among lightweight OCR VLMs. For capability, we propose Agentic Data Flow, an agent-driven data construction system that transforms model weaknesses into executable data requirements and autonomously performs material search, quality verification, and pipeline development. It substantially improves long-tail capabilities in ancient-script OCR, fine-grained chart and table parsing, multi-image text-centric QA, low-resource multilingual parsing, and document hallucination evaluation. HunyuanOCR-1.5 ranks among the top-tier end-to-end OCR solutions on OmniDocBench v1.6 while achieving new performance milestones across these long-tail tasks. Combined with an upgraded pretraining and post-training recipe, HunyuanOCR-1.5 further extends its capability in high-resolution, long-context, and multi-task scenarios. Experiments demonstrate faster inference, broader OCR capability coverage, and the deployment advantages of a lightweight end-to-end model. We will release the model weights and training code to support future research and real-world OCR applications.

Problem

• lightweight OCR
• vision-language model
• long-tail tasks
• document understanding
• inference efficiency

Innovation

• DFlash
• Agentic Data Flow
• lightweight OCR VLM
• end-to-end document understanding
• long-tail OCR tasks