LlamaIndex 发布 ParseBench-2：6 维度评测 40 款 OCR 工具
做 RAG 或文档处理的都该看看：LlamaIndex 把 40 款 OCR 工具按 6 个维度拆开评测，Opus 5.5 表格强图表弱，Extend 和 Datalab 各有强项，选型时能直接对着自己场景挑。数据集也开源了。
Jerry Liu 团队发布文档解析基准 ParseBench-2，首日覆盖 40 款工具，包括专用 OCR 服务、前沿模型和开源方案，并按表格、图表、表单等 6 个维度分别打分。基准包含 4300+ 文档页（较 ParseBench-1 翻倍），为 330+ 种错误类型逐一标注严重度，让指标分数与人工判断对齐，区别于 OmniDocBench 的确定性指标和 ParseArena 的 LLM-as-a-judge 方法。
分维度看结论更细：Opus 5.5 擅长表格但图表较弱；专用服务商各有侧重，Extend 在表单上表现好，Datalab 在表格上占优。总体分只是复合平均，团队建议按自己业务关心的错误类型自定义严重度权重。数据集已在 Hugging Face 开放，排行榜托管在 parsebench.ai。
同时团队发布 OCRia-1.5 文档解析模型家族，为 LlamaParse 各模式提供支撑，在复杂表格、表单和图表解析上较上一代有明显提升。ParseBench-1 已有 5 万+ 月下载量，被 NVIDIA、Cohere、Perplexity 等公司采用。
There are thousands of OCR tools out there, and ParseBench-2 represents our attempt to not only build a standardized benchmark, but a framework for understanding which tools do better on which dimensions. On day 1, we’ve benchmarked 40 different tools across 6 different dimensions: * specialized OCR tools * frontier models * open-weight solutions We’ve aggressively tried to improve the metrics from ParseBench-1 to reflect real penalties while remaining quantitative. *** For what it’s worth, I would take the overall score with a grain of salt, since it’s a simple composite average. A lot of the insights/nuances in the benchmark come from analyzing each individual dimension and understanding the tagging behind each failure case. It’s fascinating to see where models and other vendors do well - we don’t top the charts in every dimension relative to price point! ✅ frontier models are hillclimbing certain dimensions while ignoring others. e.g. opus 5.5 is quite good at tables, while bad at charts ✅ specialized OCR providers do well on various dimensions, e.g. Extend on forms, Datalab on tables, (we also fixed reducto r-1 chart evals) There has been some feedback on annotations and metrics that we’re continuing to take into account, and will create new versions soon. Also reminder that if you care about certain errors more than others, you're welcome to customize the severities for the use cases you care about. We’re going to be adding a ton more models to the leaderboard, come check it out: parsebench.ai Hugging Face: huggingface.co/datasets/llama…
This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include ParseBench, LlamaParse, OCR, LlamaIndex.
Source attribution: curated from Jerry Liu (@jerryjliu0).
