View-to-Path on VIEWSUITE View-to-Path (Short)
72.4AccuracyGPT-5.4 Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5.4 ProModel Category=Proprietary2026.05 | 72.4 | |
| GPT-5.4Model Category=Proprietary2026.05 | 60.5 | |
| Gemini 3.1 ProModel Category=Proprietary2026.05 | 53 | |
| GPT-5.1Model Category=Proprietary2026.05 | 52.4 | |
| Claude Opus 4.6Model Category=Proprietary2026.05 | 47.6 | |
| Grok 4.20 BetaModel Category=Proprietary2026.05 | 44.9 | |
| Gemini 3 ProModel Category=Proprietary2026.05 | 44.9 | |
| Qwen3.5-397BModel Category=Open-Weight2026.05 | 44.3 | |
| Qwen3-VL-32BModel Category=Open-Weight2026.05 | 41.1 | |
| Qwen2.5-VL-72BModel Category=Open-Weight2026.05 | 35.7 | |
| GLM-4.6VModel Category=Open-Weight2026.05 | 31.4 | |
| Qwen2.5-VL-7BModel Category=Open-Weight2026.05 | 27 | |
| Random ResponseModel Category=Baseline2026.05 | 24.3 | |
| Kimi K2.5Model Category=Open-Weight2026.05 | 18.4 |