Spatial Reasoning on ReSQ
85.41GPT-grid ScoreGPT-5.1
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GPT-5.1Setting=FG (Full-grid)2026.06 | 85.41 | — | |
| Qwen3-32BSetting=PG (Pruned-grid)2026.06 | 85.25 | 79.83 | |
| GPT-5.1Setting=PG (Pruned-grid)2026.06 | 84.59 | — | |
| Qwen3-32BSetting=FG (Full-grid)2026.06 | 83.28 | 79.5 | |
| GPT-5.1Setting=Text2026.06 | 82.6 | — | |
| LLaMA3.1-70BSetting=FG (Full-grid)2026.06 | 80.33 | 78.2 | |
| Qwen3-32BSetting=Text2026.06 | 80.3 | 80.3 | |
| LLaMA3.1-70BSetting=PG (Pruned-grid)2026.06 | 80 | 78.2 | |
| GPT-5.1Setting=Rel. Trip.2026.06 | 79.18 | — | |
| LLaMA3.1-70BSetting=Text2026.06 | 77.05 | 77.05 | |
| LLaMA3.1-70BSetting=Rel. Trip.2026.06 | 76.23 | 69.84 | |
| Qwen3-32BSetting=Rel. Trip.2026.06 | 68.85 | 68.85 |