Multimodal Perception on MMMU (val)
82AccuracyGemini 2.5 Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini 2.5 Pro2026.02 | 82 | |
| Gemini 2.5 Flash2026.02 | 79.7 | |
| Interfaze-Beta2026.02 | 77.33 | |
| Claude Opus 4Reasoning=Thinking2026.02 | 76.5 | |
| GPT-4.12026.02 | 74.8 | |
| Claude Sonnet 4Reasoning=Thinking2026.02 | 74.4 |