Skip to content

Benchmark

OmniDocBench · NED

Lower scores are better.

Models measured
8
Scores recorded
10
Sources
1
IQ weight
No weight published

Score distribution

0.16
Mean
0.14
Median
0.054
Std dev
0.115
10th percentile
0.242
90th percentile
0.109
Min
0.1090.269
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1GPT 5.4no reasoning· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.109
2Gemini 3 Pro PreviewGemini 3 Pro Model Card · results as of November 2025 · overall edit distance · lower is better· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.115
3Gemini 3 FlashOmniDocBench 1.5 · overall edit distance · lower is better· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.121
4GPT-5.4 minixhigh · OmniDocBench 1.5 · no tools · reasoning_effort none · overall edit distance· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.126
5Gemma 4 31B ITGemma 4 Technical Report · Table 6 · thinking · 1120 vision tokens · lower is better· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.131
6GPT 5.2no reasoning· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.14
7Gemma 4 26B A4BGemma 4 Technical Report · Table 6 · thinking · 1120 vision tokens · lower is better· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.149
8Gemma 4 31B ITGemma 4 Technical Report · Table 12 · thinking · 280 vision tokens · lower is better· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.201
9GPT 5.4 Nanoxhigh · OmniDocBench 1.5 · no tools · reasoning_effort none · overall edit distance· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.242
10Gemma 4 26B A4BGemma 4 Technical Report · Table 12 · thinking · 280 vision tokens · lower is better· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.269

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.