List your company

Open-Ended Question Answering on Earth Observation

97.05Judge ScoreQwen3
90.75892.391594.02595.6585Mar 20, 2026

Evaluation results

MethodLinks
Qwen3Size (B)=235-A222026.0397.0548.242.17
GPT-4.1Size (B)=1800*2026.0396.4849.772.83
Mistral Medium 3.1Size (B)=200*2026.0396.4550.24.17
EVE-InstructSize (B)=242026.0396.4—3.5
GPT OSSSize (B)=120A52026.0394.250.34.83
GPT-5 nanoSize (B)=20*2026.0392.250.25.33
MiniMax m2.5Size (B)=230A102026.039151.15.17