Spatial Understanding on SAT
88ScoreGemini
Evaluation Results
| Method | Links | |
|---|---|---|
| GeminiModel Version/Scale=3.0 Pro2026.04 | 88 | |
| HY-EmbodiedModel Version/Scale=0.5 MoE A32B2026.04 | 87.3 | |
| SeedModel Version/Scale=2.02026.04 | 86.2 | |
| QwenModel Version/Scale=3.5 A17B2026.04 | 86 | |
| Gemini2.5-ProParams=–2025.11 | 79.33 | |
| KimiModel Version/Scale=K2.52026.04 | 79.3 | |
| MiMo-EmbodiedNumber of Parameters=7B2026.04 | 78.7 | |
| MiMo-EmbodiedParams=7B2025.11 | 78.67 | |
| HY-EmbodiedModel Version=0.5 MoT-2B, Thinking Mode=thinking2026.04 | 76.7 | |
| Claude-Sonnet-4Params=–2025.11 | 75.33 | |
| RoboBrain-2.0Params=7B2025.11 | 75.33 | |
| MagmaParams=8B2025.11 | 71.33 | |
| GPT-4oParams=–2025.11 | 66.67 | |
| Cosmos-Reason1Params=7B2025.11 | 60.67 | |
| MiMo-VLParams=7B2025.11 | 59.33 | |
| InternVL3.5Params=8B2025.11 | 59.33 | |
| RoboBrain-1.0Params=7B2025.11 | 59.33 | |
| VeBrainParams=8B2025.11 | 58 | |
| QwenModel Version=3-VL, Number of Parameters=4B2026.04 | 56.7 | |
| Qwen-VL-MaxParams=–2025.11 | 56.67 | |
| Qwen2.5-VLParams=7B2025.11 | 52 | |
| RoboBrainModel Version=2.5, Number of Parameters=4B2026.04 | 51.3 | |
| QwenModel Version=3-VL, Number of Parameters=2B2026.04 | 45.3 |