LlamaIndex's ExtractBench Tests 14 Frontier Systems on 370 Enterprise Documents
llama_index · x · 2026-08-26
LlamaIndex released ExtractBench, a benchmark for schema-guided document extraction, with a technical walkthrough livestream by co-founder and CTO Simon Suo on Aug 26.
- Covers 370 enterprise documents, 67 document types, 4,800+ pages
- Evaluates 14 frontier systems across VLMs, coding agents, and specialized extraction APIs
- Focuses on real-world hard cases: scanned forms, nested tables, 40-page financial reports with merged headers
- The walkthrough covers what makes extraction hard, where existing benchmarks miss it, and cost-vs-accuracy tradeoffs
Related event: LlamaIndex Launches ExtractBench; Qwen Tops Document Extraction(3 posts)→
More from Research
- SHOES: 10 minutes of real-robot RL lets quadrupeds walk in shoes, CoRL 2026 paper — kscottz · 2026-10-16
- Post-trained NVIDIA Nemotron-3.5-Super-VL hits 35% on AutoResearchExam, matching a 20× larger model — AlexGDimakis · 2026-10-16
- U of T team mines 122,194 patient reviews across 45 Ontario hospitals with NLP to expose satisfaction and equity patterns — ishtiaqueSIA · 2026-10-16
- University of Toronto hiring AI postdoc: builders of interventions, not just benchmarkers, due Oct 30 — ishtiaqueSIA · 2026-10-16
- WorldMemArena Accepted at NeurIPS: Memory Storage ≠ Memory Use — xwang_lk · 2026-10-16
- Mech Interp Veteran Rethinks Alignment, Misgeneralization, CoT Oversight — akbirthko · 2026-10-16