LLM Stats (OmniDocBench 1.5): leaderboard
Evaluates vision-language models on diverse real-world document understanding tasks spanning tables, charts, formulas, and multi-page layouts across multiple languages.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Not forecast. 20 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | MiniMax-M3 | 91.6 | #136 |
| 2 | Qwen 3.7 Plus | 91.4 | #92 |
| 3 | Qwen 3.6 Plus | 91.2 | #120 |
| 4 | Qwen 3.8 27B | 91.1 | #105 |
| 5 | Qwen 3.6 35B A3B | 89.9 | #226 |
| 6 | Qwen 3.5 122B A10B | 89.8 | #173 |
| 7 | Qwen 3.5 35B A3B | 89.3 | #256 |
| 8 | GPT-5.4 | 89.1 | #79 |
| 9 | Qwen 3.5 27B | 88.9 | #187 |
| 10 | Kimi K2.5 | 88.8 | #144 |
| 11 | GPT-5.5 Instant | 87.5 | #129 |
| 12 | GPT-5.4 Mini | 87.37 | #212 |
| 13 | GPT-5.4 Nano | 75.81 | #328 |
| 14 | Muse Glimmer 30B | 75.8 | #178 |
| 15 | DiffusionGemma 26B-A4B | 31.9 | #698 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-omnidocbench-1-5 · How It Works · Data refreshed daily, snapshot 2026-10-20.