View-to-Path on VIEWSUITE View-to-Path (Long)
47.7AccuracyGemini 3.1 Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini 3.1 ProModel Category=Proprietary2026.05 | 47.7 | |
| Grok 4.20 BetaModel Category=Proprietary2026.05 | 44.5 | |
| GPT-5.4 ProModel Category=Proprietary2026.05 | 39 | |
| Claude Opus 4.6Model Category=Proprietary2026.05 | 38.4 | |
| GPT-5.4Model Category=Proprietary2026.05 | 37.5 | |
| Gemini 3 ProModel Category=Proprietary2026.05 | 35.5 | |
| GPT-5.1Model Category=Proprietary2026.05 | 33.4 | |
| Qwen3.5-397BModel Category=Open-Weight2026.05 | 30.8 | |
| Qwen2.5-VL-72BModel Category=Open-Weight2026.05 | 29.9 | |
| GLM-4.6VModel Category=Open-Weight2026.05 | 29.7 | |
| Kimi K2.5Model Category=Open-Weight2026.05 | 29.4 | |
| Qwen3-VL-32BModel Category=Open-Weight2026.05 | 28.5 | |
| Random ResponseModel Category=Baseline2026.05 | 26.5 | |
| Qwen2.5-VL-7BModel Category=Open-Weight2026.05 | 22.7 |