True-False Question Answering on Image Implication True-False 1.0 (test)
74AccuracyMetaphorStar-32B
Evaluation Results
| Method | Links | |
|---|---|---|
| MetaphorStar-32BParameters=32B2026.02 | 74 | |
| MetaphorStar-7BParameters=7B2026.02 | 70 | |
| Doubao-1.5-thinking-vision-pro2026.02 | 62 | |
| MetaphorStar-3BParameters=3B2026.02 | 62 | |
| Gemini-3.0-pro2026.02 | 58 | |
| QwenVL-2.5-32BParameters=32B2026.02 | 56 | |
| Gemini-2.5-flash2026.02 | 56 | |
| Gemini-2.5-flash-thinking2026.02 | 54 | |
| Claude-4.0-Sonnet2026.02 | 52 | |
| QwenVL-2.5-72BParameters=72B2026.02 | 50 | |
| GPT-4o2026.02 | 50 | |
| GLM-4V-plus2026.02 | 42 | |
| o4-mini2026.02 | 42 | |
| GPT-4.12026.02 | 40 | |
| GLM-4.1V-8BParameters=8B2026.02 | 38 | |
| Claude-3.5-Sonnet2026.02 | 38 | |
| GPT-4o-mini2026.02 | 36 | |
| InternVL3-78BParameters=78B2026.02 | 36 | |
| Grok-32026.02 | 36 | |
| Grok-3-reasoning2026.02 | 36 | |
| QwenVL-2.5-7BParameters=7B2026.02 | 28 | |
| QVQ-72BParameters=72B2026.02 | 28 | |
| QwenVL-2.5-3BParameters=3B2026.02 | 20 | |
| DeepSeek-VL22026.02 | 20 | |
| LLaVA-1.5-7BParameters=7B2026.02 | 0 |