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

Open-ended Question Answering on OlymBench Phys v1 (test)

53.9Problem Level Score

Qwen3-VL-32B-Thinking

Updated 19d ago

Evaluation Results

Method	Links
Qwen3-VL-32B-Thinking 2026.05		53.9
Claude Sonnet 4.5 2026.05		50.4
Physics-R1 (binary, 3-seed mean ±σ) 2026.05		46.2
Physics-R1 (binary, seed 42) 2026.05		45.4
Physics-R1 (dense) 2026.05		40.5
Qwen3-VL-8B-Thinking (base) 2026.05		39.3
Gemini 2.5 Pro 2026.05		37.4
GPT-4o 2026.05		19.7
InternVL3-8B 2026.05		10.4