MMM-Bench - Level 2 (Business Object): leaderboard
Metric: Macro-F1 (%) at taxonomy level 2 (Business Object), zero-shot hierarchical classification of the test split (20 percent of 5,990 real Alibaba business documents in 12 domains, mostly Chinese) from OCR text, page images and filename against a flattened five-level taxonomy, temperature 0.1; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.6 | 81.89 |
| 2 | Gemini 3.1 Pro (Preview) | 79.94 |
| 3 | Claude Opus 4.5 | 79.67 |
| 4 | GPT-5.4 | 78.11 |
| 5 | Qwen 3.6 Plus | 77.38 |
| 6 | Kimi K2.5 | 76.62 |
| 7 | MiniMax-M2.7 | 76.14 |
| 8 | Qwen 3.5 27B | 76.1 |
| 9 | GPT-5.2 | 75.97 |
| 10 | GLM-5 | 74.23 |
| 11 | Qwen 3 VL 235B A22B | 71.83 |
| 12 | Qwen 3.5 35B A3B | 70.4 |
| 13 | Qwen 3.5 9B | 67.09 |
| 14 | Qwen 2.5 VL 7B | 18.75 |
Interactive version: theaggregate.ai/benchmark?slug=mmm-bench-level-2-business-object · How It Works · Data refreshed daily, snapshot 2026-10-07.