Robot Spatial Reasoning on ROBOSPATIAL
76.3AccuracyGPT-5-mini
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5-miniGrounded=false2025.12 | 76.3 | |
| GPT-5-miniGrounded=true2025.12 | 76.2 | |
| VALOR2025.12 | 69.5 | |
| VALORModel Category=Open-Source2025.12 | 69.5 | |
| VALOR-RLModel Category=Open-Source2025.12 | 61.8 | |
| Gemini-2.0-FlashModel Category=Proprietary2025.12 | 60.9 | |
| Qwen2.5-VL-7BModel Category=Open-Source2025.12 | 60.1 | |
| Claude-3.5-HaikuModel Category=Proprietary2025.12 | 59.6 | |
| GPT-4oModel Category=Proprietary2025.12 | 57.9 | |
| Llama3.2-11BModel Category=Open-Source2025.12 | 39.9 |