LlamaIndex Releases ExtractBench, VLMs Fail at Attribution

LlamaIndex released ExtractBench, a benchmark for extracting information from complex enterprise documents that requires correct source attribution. VLMs and coding agents showed near-zero attribution ability, struggling with long documents and hallucinations.

2026-08-16 ~ 2026-08-17 · 2 related posts

• LlamaIndex releases ExtractBench for complex enterprise document information extraction — RexDouglass · 2026-08-16
• ExtractBench Reveals VLMs and Coding Agents Fail at Document Grounding — llama_index · 2026-08-17