Visual Imagination Reasoning on Hyperphantasia Medium
100Accuracy: 7 Segments (Interp)Human
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Human2026.02 | 100 | 94 | 91.33 | 54.4 | 84.93 | |
| Gemini 3 pro preview2026.02 | 96 | 86 | 38 | 32 | 63 | |
| o4-mini2026.02 | 85 | 69 | 26 | 25 | 51.25 | |
| Gemini 2.5 pro2026.02 | 44 | 74 | 26 | 24 | 42 | |
| GPT4-o2026.02 | 4 | 80 | 23 | 17 | 31 | |
| Monet2026.02 | 1 | 35 | 23 | 27 | 21.5 | |
| Claude 3.7 Sonnet2026.02 | 0 | 56 | 19 | 21 | 24 | |
| Qwen VL 2.5 7B2026.02 | 0 | 36 | 18 | 22 | 19 | |
| CapImagine2026.02 | 0 | 44 | 23 | 29 | 24 | |
| Random Guess2026.02 | 0 | 25 | 25 | 25 | 25 |