Earth-science reasoning on MSEarthMCQ
61.8AccuracyGPT-5
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5Model Type=Closed-Source2026.05 | 61.8 | |
| Gemini-2.5-ProModel Type=Closed-Source2026.05 | 55.8 | |
| Kimi-K2.5Model Type=Open-Source & Baselines2026.05 | 55 | |
| GLM-4.6VModel Type=Open-Source & Baselines2026.05 | 54.2 | |
| Gemini-2.5-FlashModel Type=Closed-Source2026.05 | 54 | |
| MAESTROModel Type=Ours2026.05 | 52.4 | |
| Qwen3-VL-32BModel Type=Open-Source & Baselines2026.05 | 51 | |
| Visual-ARFTModel Type=Think with Images Methods2026.05 | 46.6 | |
| DeepEyes-v2Model Type=Think with Images Methods2026.05 | 46.4 | |
| PixelReasonerModel Type=Think with Images Methods2026.05 | 46.2 | |
| VisionReasonerModel Type=Think with Images Methods2026.05 | 45.8 | |
| VTOOL-R1Model Type=Think with Images Methods2026.05 | 45.4 | |
| VTS-VModel Type=Think with Images Methods2026.05 | 45.4 | |
| DeepEyesModel Type=Think with Images Methods2026.05 | 45 | |
| Chain-of-FocusModel Type=Think with Images Methods2026.05 | 44.2 | |
| MathCoder-VLModel Type=Think with Images Methods2026.05 | 43.8 | |
| ThymeModel Type=Think with Images Methods2026.05 | 42.2 | |
| GPT-4oModel Type=Closed-Source2026.05 | 42 | |
| Direct AnsweringModel Type=Open-Source & Baselines2026.05 | 39.8 | |
| Untrained ModelModel Type=Open-Source & Baselines2026.05 | 36.8 |