Share your thoughts, 1 month free Claude Pro on usSee more

Earth-science reasoning on MSEarthMCQ

61.8Accuracy

GPT-5

Updated 12d ago

Evaluation Results

Method	Links
GPT-5 2026.05		61.8
Gemini-2.5-Pro 2026.05		55.8
Kimi-K2.5 2026.05		55
GLM-4.6V 2026.05		54.2
Gemini-2.5-Flash 2026.05		54
MAESTRO 2026.05		52.4
Qwen3-VL-32B 2026.05		51
Visual-ARFT 2026.05		46.6
DeepEyes-v2 2026.05		46.4
PixelReasoner 2026.05		46.2
VisionReasoner 2026.05		45.8
VTOOL-R1 2026.05		45.4
VTS-V 2026.05		45.4
DeepEyes 2026.05		45
Chain-of-Focus 2026.05		44.2
MathCoder-VL 2026.05		43.8
Thyme 2026.05		42.2
GPT-4o 2026.05		42
Direct Answering 2026.05		39.8
Untrained Model 2026.05		36.8