Question Answering on VERHallu
94.9QA-Causal ScoreHuman
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Human2026.01 | 94.9 | 92.4 | 93.8 | |
| Gemini-3-Pro2026.01 | 65.7 | 55 | 40 | |
| ChatGPT-4o2026.01 | 46.5 | 36 | 42.5 | |
| Qwen-VL-2.5-32BParameters=32B2026.01 | 39.6 | 34.4 | 31.4 | |
| Qwen-VL-2.5-72BParameters=72B2026.01 | 28.9 | 28.7 | 34.8 |