做文档解析、RAG 或 AI Agent 的团队终于有了一个靠谱的评测标准——ParseBench 覆盖了企业文档的真实痛点，建议直接拿去测你的模型或产品。
LlamaIndex 在 CVPR 2026 上发布了 ParseBench，这是目前最全面的文档理解基准测试，专门用于评估视觉语言模型（VLM）对真实企业文档的解析能力。该基准包含 2000 页真实企业文档、167K+ 测试规则，覆盖表格、图表、视觉定位、语义格式和内容忠实度五个维度。核心目标是衡量模型能否正确语义理解文档，避免过拟合到特定基准。当前前沿模型更擅长编程、数学和科学推理，而文档 OCR 的 100% 准确解析仍是最终挑战，ParseBench 旨在推动这一方向进步。
We're presenting ParseBench at CVPR 2026! ParseBench is the most comprehensive document understand...
We're presenting ParseBench at CVPR 2026! ParseBench is the most comprehensive document understanding benchmark for VLMs. ✅ It contains 2k pages of real-world enterprise documents ✅ It has comprehensive evaluation metrics around tables, charts, visual grounding, semantic formatting, and content faithfulness The core goal is measuring whether models can semantically interpret a document in the right way, without having models overfit to our precise benchmark. Parsing 100% of PDFs to 100% accuracy is the final boss for document OCR. In general, the latest frontier models have been tuned for coding, math, and scientific reasoning as opposed to precise visual understanding; hope more benchmarks that these will encourage overall progress towards solving this problem! Poster is below. If you want to learn more come check out our site or 30-page ArXiv paper: ParseBench: parsebench.ai ArXiv: arxiv.org/abs/2604.08538 LlamaIndex 🦙 @llama_index We're presenting ParseBench at CVPR 2026 today. 🦙 Come learn why document understanding is an AGI-complete problem (an agent can't act on a doc it can't correctly read, and reading a real enterprise table is harder than it looks). The first doc-parsing benchmark built for AI agents: 2,000+ human-verified pages 167K+ test rules 5 dimensions: tables, charts, faithfulness, formatting, grounding Fully open source. 📍 Talk TODAY, June 4, 9–10 AM at CVPR. Come say hi 👇 huggingface.co/datasets/llama… GVT github.com/run-llama/Pars… TWY arxiv.org/abs/2604.08538 b48oJl 🔗 View Quoted Tweet 💬 2 🔄 2 ❤️ 7 👀 472 📊 4 ⚡