ParseBench
Progress Over Time
Interactive timeline showing model performance evolution on ParseBench
ParseBench Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 |  Cohere | 2B | 8K | — |
What is ParseBench?
ParseBench evaluates document parsing for AI-agent workflows across enterprise pages (insurance, finance, government). It measures semantic correctness rather than text similarity, with capability dimensions including tables, charts, content faithfulness, semantic formatting, and visual grounding. Vendor Parse comparisons often report a three-dimension average (tables, content faithfulness, semantic formatting) under Aug 2026 evaluation rules.
ParseBench is a multimodal benchmark evaluating models on multimodal, image to text, and vision tasks. LLM Stats tracks 1 models on this benchmark, scored on a 0–1 scale. The current average is 0.8, with the leader at 0.8.
Compare leaders on the best AI for multimodal, best AI for image to text and best AI for vision leaderboards.
Current leaders
Parse from Cohere currently leads the ParseBench leaderboard with a score of 0.792 across 1 evaluated AI models.
Source paper
- Title
- ParseBench: A Document Parsing Benchmark for AI Agents
- Authors
- Boyang Zhang, Sebastián G. Acosta, Preston Carlson, Sacha Bron, and 3 others
- Published
- arXiv
- 2604.08538
Abstract
AI agents are changing the requirements for document parsing. What matters is semantic correctness: parsed output must preserve the structure and meaning needed for autonomous decisions, including correct table structure, precise chart data, semantically meaningful formatting, and visual grounding. Existing benchmarks do not fully capture this setting for enterprise automation, relying on narrow document distributions and text-similarity metrics that miss agent-critical failures. We introduce ParseBench, a benchmark of ${\sim}2{,}000$ human-verified pages from enterprise documents spanning insurance, finance, and government, organized around five capability dimensions: tables, charts, content faithfulness, semantic formatting, and visual grounding. Across 14 methods spanning vision-language models, specialized document parsers, and LlamaParse, the benchmark reveals a fragmented capability landscape: no method is consistently strong across all five dimensions. LlamaParse Agentic achieves the highest overall score at 84.9%, and the benchmark highlights the remaining capability gaps across current systems. Dataset and evaluation code are available on https://huggingface.co/datasets/llamaindex/ParseBench and https://github.com/run-llama/ParseBench.
FAQ
Common questions about the ParseBench benchmark and leaderboard.