Segment Intrusion on Segment Intrusion Restaurant domain 1.0 (test)
98SI-EHuman Performance
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Human Performance2026.05 | 98 | 97 | 97 | 87 | 94.75 | |
| o4-mini2026.05 | 98 | 83.71 | 95 | 77.18 | 88.47 | |
| o3-mini2026.05 | 97.5 | 82.12 | 94 | 74.63 | 87.06 | |
| gemini-2.5-flash-preview-04-172026.05 | 97.5 | 82.5 | 76.27 | 63.98 | 80.06 | |
| claude-3.7-sonnet-202502192026.05 | 96.76 | 77.5 | 95 | 72.99 | 85.56 | |
| llama-3.3-70b-instruct-turbo2026.05 | 95.52 | 74.19 | 88.78 | 64.89 | 80.84 | |
| deepseek-v32026.05 | 95 | 78.5 | 85.57 | 70.22 | 82.32 | |
| gpt-4o2026.05 | 93.5 | 79.31 | 89.28 | 73.48 | 83.89 | |
| qwen2.5-72b-instruct-turbo2026.05 | 91.27 | 74 | 91.04 | 65.86 | 80.54 | |
| claude-3.5-haiku-202410222026.05 | 90.51 | 72.55 | 82 | 66.34 | 77.85 | |
| gemini-2.0-flash2026.05 | 90 | 66.5 | 86.28 | 63.77 | 76.64 | |
| gemini-2.0-flash-lite2026.05 | 79.82 | 62.57 | 76.24 | 66.67 | 71.32 | |
| qwen2.5-7b-instruct-turbo2026.05 | 74.81 | 53.5 | 40.1 | 31.75 | 50.04 | |
| llama-4-maverick-17b-128e-instruct-fp8Precision=fp82026.05 | 55.94 | 45.59 | 21.48 | 19.82 | 35.71 | |
| llama-3.2-3b-instruct-turbo2026.05 | 27.07 | 20.76 | 10.49 | 7.5 | 16.45 |