Document understanding benchmark measured by edit distance on complex extraction and grounded document tasks.
As of March 2026, GPT-5.4 leads the OmniDocBench 1.5 leaderboard with 0.1090 , followed by GPT-5.4 mini (0.1263) and GPT-5 mini (0.1791).
GPT-5.4
OpenAI
GPT-5.4 mini
OpenAI
GPT-5 mini
OpenAI
According to BenchLM.ai, GPT-5.4 leads the OmniDocBench 1.5 benchmark with a score of 0.1090, followed by GPT-5.4 mini (0.1263) and GPT-5 mini (0.1791). The top models are clustered within 0.0701 points, suggesting this benchmark is nearing saturation for frontier models.
4 models have been evaluated on OmniDocBench 1.5. The benchmark falls in the Multimodal & Grounded category, which carries a 12% weight in BenchLM.ai's overall scoring system. OmniDocBench 1.5 is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
Year
2026
Tasks
Document understanding tasks
Format
Edit-distance error metric
Difficulty
Grounded document reasoning
OmniDocBench is an error metric where lower scores indicate better document understanding and extraction quality.
Introducing GPT-5.4 mini and nanoDocument understanding benchmark measured by edit distance on complex extraction and grounded document tasks.
GPT-5.4 by OpenAI currently leads with a score of 0.1090 on OmniDocBench 1.5.
4 AI models have been evaluated on OmniDocBench 1.5 on BenchLM.
Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.
Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.