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

Scientific Multimodal Reasoning on VRSBench

74.32Accuracy

S1-VL-32B-RL

Updated 1mo ago

Evaluation Results

Method	Links
S1-VL-32B-RL 2026.04		74.32
S1-VL-32B-SFT 2026.04		72.34
Qwen3-VL-235B-A22B-Thinking 2026.04		68.94
Qwen3-VL-32B-Thinking 2026.04		68.41
GPT-5 2026.04		65.89
Gemini 2.5 Pro 2026.04		65.7
Gemini 2.5 Flash 2026.04		64.23
Intern-S1 2026.04		63.48
Thyme-VL 2026.04		57.61
Intern-S1-mini 2026.04		54.89