List your company

Physical Reasoning on WorldModelBench

41.8General ScoreQwen9B-base
-0.63210.38421.432.416May 11, 2026

Evaluation results

MethodLinks
Qwen9B-baseCoT=true2026.0541.841.641.747.8
Qwen27B-baseCoT=true2026.0537.346.942.148.3
Qwen9B-baseCoT=false2026.0528.23531.637.7
Claude (Opus 4.7)CoT=true2026.0525.234.429.834.3
Qwen27B-baseCoT=false2026.0521.340.931.135.7
Gemini-3.1-ProCoT=false2026.0517.315.916.6-6.3
Gemini-3.1-ProCoT=true2026.0517.217.417.3-6.5
Claude (Opus 4.7)CoT=false2026.0516.21013.114.5
PhyJudge-9BCoT=false2026.0515.53.3-1.1