Open-Style Question Answering on Image Implication Open-Style 1.0 (test)
3.94ScoreMetaphorStar-32B
Evaluation Results
| Method | Links | |
|---|---|---|
| MetaphorStar-32BParameters=32B2026.02 | 3.94 | |
| Gemini-3.0-pro2026.02 | 3.82 | |
| Claude-4.0-Sonnet2026.02 | 3.46 | |
| InternVL3-78BParameters=78B2026.02 | 3.42 | |
| Gemini-2.5-flash-thinking2026.02 | 3.42 | |
| Gemini-2.5-flash2026.02 | 3.34 | |
| GPT-4.12026.02 | 3.3 | |
| o4-mini2026.02 | 3.26 | |
| Grok-32026.02 | 3.24 | |
| Claude-3.5-Sonnet2026.02 | 3.22 | |
| MetaphorStar-7BParameters=7B2026.02 | 3.22 | |
| Doubao-1.5-thinking-vision-pro2026.02 | 3.16 | |
| QVQ-72BParameters=72B2026.02 | 3.1 | |
| QwenVL-2.5-32BParameters=32B2026.02 | 3.08 | |
| Grok-3-reasoning2026.02 | 3.06 | |
| MetaphorStar-3BParameters=3B2026.02 | 3.06 | |
| GLM-4V-plus2026.02 | 3.01 | |
| GPT-4o-mini2026.02 | 2.98 | |
| GPT-4o2026.02 | 2.94 | |
| DeepSeek-VL22026.02 | 2.82 | |
| GLM-4.1V-8BParameters=8B2026.02 | 2.6 | |
| QwenVL-2.5-3BParameters=3B2026.02 | 2.44 | |
| QwenVL-2.5-7BParameters=7B2026.02 | 2.34 | |
| LLaVA-1.5-7BParameters=7B2026.02 | 2.06 | |
| QwenVL-2.5-72BParameters=72B2026.02 | 1.56 |