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

Scientific Multimodal Reasoning on Galaxy-10

57.72Accuracy

GPT-5

Updated 1mo ago

Evaluation Results

Method	Links
GPT-5 2026.04		57.72
S1-VL-32B-RL 2026.04		51.52
Intern-S1 2026.04		48.37
S1-VL-32B-SFT 2026.04		42.45
Intern-S1-mini 2026.04		39.06
Qwen3-VL-235B-A22B-Thinking 2026.04		33.93
Qwen3-VL-32B-Thinking 2026.04		30.44
Thyme-VL 2026.04		27.84
Gemini 2.5 Pro 2026.04		25.14
Gemini 2.5 Flash 2026.04		15.78