Paper Classification on marine Pb benchmark
96AccuracyGemini-2.5-pro
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Gemini-2.5-proSize=N/A2026.05 | 96 | 91.3 | |
| CompassSize=8B2026.05 | 96 | 94.2 | |
| CompassSize=32B2026.05 | 96 | 95.6 | |
| GPT-4oSize=N/A2026.05 | 93 | 84.5 | |
| Qwen2.5Size=32B2026.05 | 92 | 88.7 | |
| Qwen3Size=8B2026.05 | 88 | 80.1 | |
| Llama-3.1Size=8B2026.05 | 82 | 75.7 | |
| OceanGPTSize=7B2026.05 | 80 | 74.8 | |
| K2 (LLaMA)Size=7B2026.05 | 5 | 3.7 |