| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Multimodal Reasoning | M3-Bench web | ME Score46.4 | 24 | |
| Multimodal Reasoning | M3-Bench robot | ME Score35.7 | 24 | |
| Text-to-Code | M3²Bench | Execution Rate (%)91.4 | 14 | |
| Diagram Editing | M3²Bench | Execution Rate86.5 | 14 | |
| Diagram-to-Code | M3²Bench | Execution Rate94.3 | 14 | |
| Video Question Answering | M3-Bench web | ME55.3 | 9 | |
| Video Question Answering | M3-Bench robot | ME Score43.7 | 9 | |
| Long-video understanding | M3-Bench web | MDR55.4 | 8 | |
| Long-video understanding | M3-Bench robot | MDR47.8 | 8 |