OmniDocBench v1.6
PulseAugur coverage of OmniDocBench v1.6 — every cluster mentioning OmniDocBench v1.6 across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
New Omni-Modal Parser Youtu-Parsing-Omni Handles Text, Image, Audio, Video
Researchers have introduced Youtu-Parsing-Omni, an omni-modal parser capable of converting documents, images, audio, and video into structured, spatio-temporal records. This system utilizes a single encoder for each mod…
-
Youtu-Parsing-Omni Unifies Multi-Modal Parsing with Single Encoder
Researchers have introduced Youtu-Parsing-Omni, a unified multi-modal parser capable of processing documents, images, audio, and video into structured spatiotemporal records. This model utilizes a single encoder for eac…
-
llama.cpp updates, federated AI agents, and new OCR model released
The llama.cpp project has released version b12015, which includes support for various platforms and configurations, including openEuler with ACL Graph. Separately, a research paper titled "Federated Recommender Agents" …
-
OpenBMB releases MiniCPM-V 4.7, a 35B MoE model runnable on iPhones
OpenBMB and Face++ have released MiniCPM-V 4.7, an open-source multimodal model capable of running on mobile devices with as little as 3GB of RAM. The model comes in two versions: a 35B parameter Mixture-of-Experts (MoE…
-
New GAD-RL method boosts OCR faithfulness in vision-language models
Researchers have developed GAD-RL, a novel method to improve the faithfulness of optical character recognition (OCR) in vision-language models. This technique adaptively regulates teacher supervision during post-trainin…
-
PolyOCR introduces unified foundation models for advanced OCR tasks
Researchers have introduced PolyOCR, a family of unified OCR foundation models designed to handle a range of tasks including recognition, localization, parsing, and reasoning over text in images. These models utilize a …
-
WeVisDoc framework boosts document parsing accuracy
Researchers have introduced WeVisDoc, a novel two-stage framework designed to enhance the robustness of end-to-end document parsing models. The first stage broadens the model's coverage across various semantic, structur…
-
Jina-OCR-v1 model offers efficient document parsing on low-budget GPUs
Researchers have introduced Jina-OCR-v1, a new document parsing model designed for efficient operation on low-budget GPUs. This model integrates the DeepSeek-OCR architecture with a FastMTP speculative decoding head, wh…
-
NaviDC-OCR framework enhances document parsing with deformation-aware learning
Researchers have introduced NaviDC-OCR, a novel framework designed to enhance document parsing across both digital and camera-captured documents. This system addresses limitations in existing methods by incorporating de…
-
NaviDC-OCR framework enhances document parsing with deformation-aware learning
NaviDC-OCR is a new vision-language framework designed to improve document parsing by addressing challenges in handling distorted camera-captured documents and enhancing structural reasoning. The framework incorporates …
-
PaddlePaddle releases HPD-Parsing model with record throughput · 2 sources tracked
PaddlePaddle has released HPD-Parsing, a new lightweight document parsing model that utilizes a Hierarchical Parallel Decoding paradigm. This model achieves a new state-of-the-art score of 94.91% on the OmniDocBench v1.…
-
OvisOCR2 document parsing model achieves state-of-the-art on benchmarks · 3 sources tracked
Researchers have introduced OvisOCR2, a new 0.8 billion parameter document parsing model capable of generating Markdown representations of documents, including text, formulas, and tables, in natural reading order. The m…
-
OvisOCR2: A new 0.8B OCR model converts documents to structured Markdown
OvisOCR2 is a new 0.8B parameter end-to-end OCR model that converts full document pages into structured Markdown. Based on Qwen3.5-0.8B, it reportedly achieves high scores on document understanding benchmarks like OmniD…
-
HunyuanOCR-1.5 enhances lightweight OCR VLMs with faster inference and improved capabilities · 3 sources tracked
Researchers have introduced HunyuanOCR-1.5, an enhanced lightweight vision-language model specifically designed for Optical Character Recognition (OCR). This model unifies various document processing tasks, including pa…
-
US power sector M&A hits record $200B amid AI data center boom; Baidu's OCR model goes open-source
The US power and utility sector is experiencing a surge in M&A activity, with deal values exceeding $200 billion in the first five months of the year, driven by the demand for energy infrastructure to support data cente…
-
Baidu's Unlimited OCR model tops leaderboards and sees rapid adoption
Baidu has released and open-sourced its end-to-end OCR model, Unlimited OCR. The model quickly achieved top rankings on GitHub and Hugging Face, demonstrating strong performance in long document parsing. Unlimited OCR a…
-
Baidu's PaddleOCR-VL-1.6 sets new SOTA in document parsing
Baidu's Wenxin has released PaddleOCR-VL-1.6, a new version of its open-source OCR tool. This update achieves over 96.33% accuracy on the OmniDocBench v1.6 benchmark, surpassing major models like Gemini-3-Pro and GPT-5.…
-
New benchmarks and models advance document parsing and table extraction
Researchers have introduced new benchmarks and improved models for document parsing and table extraction. Dr. DocBench focuses on expert-level document parsing, including complex structures like chemical formulas and mu…
-
PaddleOCR-VL-1.6 sets new SOTA in document parsing
PaddlePaddle has released PaddleOCR-VL-1.6, an advanced document parsing model that achieves state-of-the-art accuracy on several benchmarks, including OmniDocBench v1.6 with a score of 96.33%. This new version incorpor…