Scene understanding on 40 harbor scenes 1.0 (test)
5.69Latency (s)gpt-4.1
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| gpt-4.1Model=gpt-4.12025.12 | 5.69 | 83 | 83 | 83 | 84 | |
| gpt-5-minimalModel=gpt-5, variant=minimal2025.12 | 7.07 | 83.3 | 82 | 82 | 86 | |
| gpt-5-lowModel=gpt-5, variant=low2025.12 | 17.21 | 84.5 | 85 | 81 | 86 | |
| gpt-5-mediumModel=gpt-5, variant=medium2025.12 | 30.04 | 86.6 | 87 | 85 | 88 | |
| gpt-5-highModel=gpt-5, variant=high2025.12 | 59.22 | 86.6 | 87 | 85 | 87 |