Visual Imagination Reasoning on Hyperphantasia Easy
100Seven Segments (Interpolation) AccHuman
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Human2026.02 | 100 | 98.86 | 100 | 100 | 99.72 | |
| Gemini 3 pro preview2026.02 | 98 | 99 | 42 | 30 | 67.25 | |
| o4-mini2026.02 | 83 | 90 | 43 | 35 | 62.75 | |
| Gemini 2.5 pro2026.02 | 51 | 97 | 31 | 26 | 51.25 | |
| GPT4-o2026.02 | 3 | 96 | 28 | 16 | 35.75 | |
| Claude 3.7 Sonnet2026.02 | 1 | 86 | 60 | 40 | 44.25 | |
| CapImagine2026.02 | 1 | 68 | 37 | 33 | 34.75 | |
| Qwen VL 2.5 7B2026.02 | 0 | 66 | 27 | 16 | 27.25 | |
| Monet2026.02 | 0 | 62 | 33 | 19 | 28.5 | |
| Random Guess2026.02 | 0 | 25 | 25 | 25 | 25 |