STEM & Reasoning on VisualPuzzle
71.48AccuracyGemini 3-Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini 3-Pro2026.02 | 71.48 | |
| ERNIE 5.02026.02 | 64.82 | |
| Gemini 2.5-Pro2026.02 | 61.51 | |
| GPT-5tier=High2026.02 | 57.75 | |
| Qwen3-VLmode=Thinking2026.02 | 57.01 | |
| ERNIE 5.0-BaseModel type=pre-trained2026.02 | 54.24 |