Multiple Choice Question Answering (Single) on Earth Observation
96.35AccuracyEVE-Instruct
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| EVE-InstructSize (B)=242026.03 | 96.35 | 3.5 | |
| Qwen3Size (B)=235-A222026.03 | 95.16 | 2.17 | |
| Mistral Medium 3.1Size (B)=200*2026.03 | 95 | 4.17 | |
| MiniMax m2.5Size (B)=230A102026.03 | 94.95 | 5.17 | |
| GPT-4.1Size (B)=1800*2026.03 | 94.37 | 2.83 | |
| GPT-5 nanoSize (B)=20*2026.03 | 91.99 | 5.33 | |
| GPT OSSSize (B)=120A52026.03 | 89.77 | 4.83 |