Multimodal & Grounded benchmark
ParseBench leaderboard
ParseBench: document parsing for AI agents (LlamaIndex leaderboard). Every model the catalog carries a published ParseBench value for, ranked by that value.
CategoryMultimodal & Grounded
MeasureDeterministic test rules per page; five dimension scores and their mean, in percent
Tasks2,078 human-verified enterprise pages from insurance, finance and government documents
DifficultyNo method is strong on all five dimensions; charts and visual grounding separate the field
Scores document parsers and vision-language models on 2,078 human-verified enterprise pages across tables, charts, content faithfulness, semantic formatting and visual grounding, with the listed cost per page.
About ParseBench
What the benchmark measures and how its value is published, from the source’s own record.
What is ParseBench?
ParseBench (ParseBench: document parsing for AI agents (LlamaIndex leaderboard)) is a multimodal & grounded benchmark published by ParseBench: A Document Parsing Benchmark for AI Agents. Scores document parsers and vision-language models on 2,078 human-verified enterprise pages across tables, charts, content faithfulness, semantic formatting and visual grounding, with the listed cost per page. Its evaluation set is described by the source as 2,078 human-verified enterprise pages from insurance, finance and government documents. The source rates its difficulty as no method is strong on all five dimensions; charts and visual grounding separate the field.
How is ParseBench scored?
Each model's result is published by ParseBench: A Document Parsing Benchmark for AI Agents as deterministic test rules per page; five dimension scores and their mean, in percent. ModelScale prints that value to 2 decimal places, exactly as published, and never converts or rescales it. This page ranks the 38 models that carry a published value, highest value first; a model the source never scored is left out rather than ranked at zero. The source does not state whether a higher value is the better result, so neither does this page.
ModelScale
