Hallucination Detection on All detection benchmark sets average latest
83.27Precisiono4-mini
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| o4-miniCategory=Prompt Based2025.12 | 83.27 | 83.12 | 82.62 | 85.23 | |
| Qwen3-32BSize=32B, Category=Prompt Based2025.12 | 81.5 | 82.42 | 81.7 | 84.24 | |
| HDM-4B-RLSize=4B, Category=Hallucination Detection Models2025.12 | 80.39 | 81.04 | 79.65 | 82.24 | |
| Qwen3-4BSize=4B, Category=Prompt Based2025.12 | 77.89 | 76.42 | 76.74 | 80.85 | |
| GPT-4.1Category=Prompt Based2025.12 | 77.84 | 75.09 | 75.37 | 79.99 | |
| Qwen3-32B†Size=32B, Category=Prompt Based2025.12 | 77.81 | 73.8 | 72.58 | 79.51 | |
| Lynx-8BSize=8B, Category=Hallucination Detection Models2025.12 | 77.79 | 77.54 | 77.25 | 81.63 | |
| MiniCheck-7BSize=7B, Category=Hallucination Detection Models2025.12 | 72.82 | 71.66 | 68.48 | 71.34 | |
| Qwen3-4B†Size=4B, Category=Prompt Based2025.12 | 70.93 | 65.35 | 64.92 | 71.83 | |
| HDM-4B-RL†Size=4B, Category=Hallucination Detection Models2025.12 | 70.67 | 71.32 | 70.13 | 73.74 | |
| MiniCheck-FT5Size=783M, Category=Hallucination Detection Models2025.12 | 67.08 | 64.06 | 62.3 | 68.03 | |
| FactCG-DBTSize=435M, Category=Hallucination Detection Models2025.12 | 66.07 | 60.45 | 58.33 | 66.45 | |
| Alignscore-LSize=355M, Category=Hallucination Detection Models2025.12 | 65.8 | 60.65 | 59.21 | 66.88 | |
| HHEMSize=110M, Category=Hallucination Detection Models2025.12 | 65.38 | 59.99 | 57.63 | 65.48 |