Geolocation Inference on DOXBENCH
96.88VRRGPT-4.1
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| GPT-4.1Top-K evaluation=Top 3, Model Type=Vanilla2025.04 | 96.88 | 14.06 | 4.29 | 98.92 | 86 | 29 | 1,916.55 | |
| GEMINI 2.5 PROTop-K evaluation=Top 3, Model Type=MLRM2025.04 | 95.07 | 9.92 | 2.98 | 99.72 | 108 | 42 | 1,987.16 | |
| LLAMA 4 MAVERICKTop-K evaluation=Top 3, Model Type=Vanilla2025.04 | 91.85 | 174.82 | 28.49 | 67.77 | 32 | 15 | 1,253.85 | |
| LLAMA 4 MAVERICKTop-K evaluation=Top 1, Model Type=Vanilla2025.04 | 88.77 | 166.61 | 30.86 | 67.72 | 31 | 17 | 1,219.01 | |
| OPENAI O3Top-K evaluation=Top 3, Model Type=MLRM2025.04 | 87.95 | 7.44 | 2.73 | 100 | 96 | 37 | 1,912.77 | |
| QVQ-MAXTop-K evaluation=Top 3, Model Type=MLRM2025.04 | 84.8 | 32.92 | 16.15 | 92.06 | 41 | 15 | 1,455.18 | |
| GEMINI 2.5 PROTop-K evaluation=Top 1, Model Type=MLRM2025.04 | 84.53 | 14.75 | 4.63 | 97.14 | 84 | 32 | 1,701.61 | |
| GPT-4.1Top-K evaluation=Top 1, Model Type=Vanilla2025.04 | 83.48 | 15.24 | 6.07 | 98.76 | 64 | 27 | 1,647.29 | |
| OPENAI O3Top-K evaluation=Top 1, Model Type=MLRM2025.04 | 80.8 | 13.56 | 5.46 | 99.02 | 71 | 34 | 1,628.5 | |
| OPENAI O4-MINITop-K evaluation=Top 3, Model Type=MLRM2025.04 | 71.88 | 11.2 | 4.31 | 100 | 71 | 30 | 1,515.72 | |
| QVQ-MAXTop-K evaluation=Top 1, Model Type=MLRM2025.04 | 66.74 | 121.06 | 24.02 | 74.44 | 37 | 13 | 1,025.05 | |
| OPENAI O4-MINITop-K evaluation=Top 1, Model Type=MLRM2025.04 | 53.79 | 15.64 | 7.04 | 98.09 | 57 | 24 | 1,105.84 | |
| CLAUDE OPUS 4Top-K evaluation=Top 3, Model Type=MLRM2025.04 | 40.75 | 20.33 | 5.49 | 90.91 | 41 | 17 | 859.03 | |
| CLAUDE OPUS 4Top-K evaluation=Top 3, Model Type=Vanilla2025.04 | 39.65 | 21.92 | 9.16 | 93.51 | 36 | 18 | 804.2 | |
| LLAMA 4 SCOUTTop-K evaluation=Top 1, Model Type=Vanilla2025.04 | 34.36 | 129.16 | 26.32 | 70.29 | 16 | 6 | 565.58 | |
| LLAMA 4 SCOUTTop-K evaluation=Top 3, Model Type=Vanilla2025.04 | 32.38 | 33.6 | 14.46 | 87.29 | 21 | 10 | 627.2 | |
| CLAUDE SONNET 4Top-K evaluation=Top 3, Model Type=Vanilla2025.04 | 27.31 | 92.15 | 8.99 | 73.04 | 28 | 15 | 516 | |
| CLAUDE OPUS 4Top-K evaluation=Top 1, Model Type=Vanilla2025.04 | 24.01 | 145.06 | 30.04 | 60.95 | 28 | 17 | 401.24 | |
| CLAUDE SONNET 4Top-K evaluation=Top 1, Model Type=Vanilla2025.04 | 23.35 | 92.68 | 9.62 | 73.47 | 25 | 13 | 444.71 | |
| CLAUDE OPUS 4Top-K evaluation=Top 1, Model Type=MLRM2025.04 | 15.64 | 108.52 | 3.36 | 69.12 | 25 | 15 | 328.11 | |
| GPT-4OTop-K evaluation=Top 3, Model Type=Vanilla2025.04 | 13.84 | 1.24 | 0.27 | 100 | 35 | 18 | 432.47 | |
| GPT-4OTop-K evaluation=Top 1, Model Type=Vanilla2025.04 | 12.95 | 2.01 | 0.4 | 100 | 29 | 15 | 389.83 | |
| CLAUDE SONNET 4Top-K evaluation=Top 3, Model Type=MLRM2025.04 | 12.11 | 21.34 | 0.62 | 88.89 | 22 | 13 | 317 | |
| CLAUDE SONNET 4Top-K evaluation=Top 1, Model Type=MLRM2025.04 | 9.47 | 4.8 | 1 | 100 | 16 | 9 | 265.25 |