List your company

Scientific Reasoning on SciKnowEval Material

71.48Mean@16GRPO
-1.725617.279736.28555.2903May 27, 2026

Evaluation results

MethodLinks
GRPOModel Size=Qwen3-4B, Training Domain=Physics2026.0571.48
GRPOModel Size=Qwen3-8B, Training Domain=Chemistry2026.0564.18
GRPOModel Size=Qwen3-8B, Training Domain=Physics2026.0563.76
ROSDModel Size=Qwen3-4B, Training Domain=Physics2026.0562.21
ROSDModel Size=Qwen3-8B, Training Domain=Physics2026.0560.33
GRPOModel Size=Qwen3-8B, Training Domain=Biology2026.0559.04
GRPOModel Size=Qwen3-8B, Training Domain=ToolUse2026.0558.82
GRPOModel Size=Qwen3-4B, Training Domain=Biology2026.0556.16
ROSDModel Size=Qwen3-8B, Training Domain=Biology2026.0556.01
ROSDModel Size=Qwen3-8B, Training Domain=ToolUse2026.0555.87
ROSDModel Size=Qwen3-4B, Training Domain=ToolUse2026.0553.3
GRPOModel Size=Qwen3-4B, Training Domain=ToolUse2026.0552.7
ROSDModel Size=Qwen3-4B, Training Domain=Biology2026.0551.37
ROSDModel Size=Qwen3-8B, Training Domain=Chemistry2026.0550.22
GRPOModel Size=Qwen3-4B, Training Domain=Chemistry2026.0549.78
SDPOModel Size=Qwen3-8B, Training Domain=Physics2026.0549.38
SDPOModel Size=Qwen3-8B, Training Domain=ToolUse2026.0544.22
SDPOModel Size=Qwen3-4B, Training Domain=Physics2026.0543.62
SDPOModel Size=Qwen3-4B, Training Domain=Biology2026.0538.16
SDPOModel Size=Qwen3-8B, Training Domain=Biology2026.0538.1
ROSDModel Size=Qwen3-4B, Training Domain=Chemistry2026.0534.64
SDPOModel Size=Qwen3-8B, Training Domain=Chemistry2026.0522.89
SDPOModel Size=Qwen3-4B, Training Domain=Chemistry2026.0520.41
SDPOModel Size=Qwen3-4B, Training Domain=ToolUse2026.051.09