LlamaIndex 发布 ExtractBenchPro：面向真实文档抽取的高难度基准

A lot of professional work around documents can be arbitrarily challenging to automate. It not only requires deep visual perception/OCR around complex entities (diagrams/charts), but also requires a reasoning trace that can reconcile disparate sources of oftentimes contradictory information. Check out the example video below around a water loss audit. Solving this task not only requires reading the water loss diagram, but also understand the exact date and subsequent changes in the appendix. We think hard document extraction tasks as a microcosm for professional work over PDFs. That's why we launched ExtractBenchPro - a benchmark that's challenging even for the latest frontier models (e.g. Astra/Opus). Even as frontier models with max thinking models push the boundaries of accuracy, there is a massive opportunity to build specialized models that can push the frontiers of cost and latency. For more info on ExtractBench Pro, check out our blog and the website! llamaindex.ai/blog/introduci… Site: extractbenchpro.ai Your browser does not support the video tag. 🔗 View on Twitter Jerry Liu @jerryjliu0 Introducing ExtractBenchPro: the most challenging benchmark for real-world document extraction. We see a massive long tail of document complexity in terms of use cases brought on from our customers, that are completely unsolved from a price-performance standpoint. These include trial balances, well logs, canvass binders, site plan reviews, duty reconciliations, loss runs, and more. Most existing extraction benchmarks are saturated by the frontier models (our cheapest modes now score 93%+ on ExtractBench). So there is enormous economic value in building a complex extraction benchmark that reflects this long-tail of knowledge work. We spent a long time curating the most complex elements that we saw from real world extraction use cases: ✅ Document conventions: containing legends or rules where the rest of the document should be interpreted ✅ Density of internal jargon: specific definitions that must be cross-referenced within a doc ✅ Aggregation and reasoning: beyond returning exact values on a page, infer values derived from aggregating multiple sources of information ✅ Drawn relations: complex diagrams like procedures and circuits ✅ Charts inside documents: read precise points within a chart ✅ Long chart sheets: some data series like well logs cram all this data into an extremely long page Within the industries where document extraction is relevant, even small mistakes can have enormous negative impacts. Therefore our headline evaluation metric is based on perfect documents: the proportion of docs with perfect (100%) value f1 score. (Besides this we also measure value f1 and grounded f1) It’s an extremely hard task for even the latest frontier models to solve, topping out at ~58% or so in Value F1 score for Codex with Astra - which is quite good but extremely expensive (almost $1 per page!) Our LlamaParse agentic/agentic plus modes are 20-40% and cover the left side of the Pareto frontier, at price ranges that are feasible for full production use. ExtractBenchPro is available today as a website/technical blog, and also on Github and HuggingFace. Check it out! Blog: llamaindex.ai/blog/introduci… Website: extractbenchpro.ai Hugging Face: huggingface.co/datasets/llama… Github (is ExtractBench): github.com/run-llama/Extr… Your browser does not support the video tag. 🔗 View on Twitter 🔗 View Quoted Tweet 💬 8 🔄 0 ❤️ 12 👀 719 📊 8 ⚡