List your company

Logical Reasoning on Knight-and-Knaves

81.8Pass@1 Rate (4-8 Roles, ID)RLCSD
62.45667.47872.577.522Jun 10, 2026

Evaluation results

MethodLinks
RLCSDModel=Qwen3-8B2026.0681.879706574
RLCSDModel=Qwen3-4B2026.0678.673665066.9
RLSDModel=Qwen3-4B2026.0676.872634865
GRPOModel=Qwen3-8B2026.0676.875634966
RLSDModel=Qwen3-8B2026.0676.677645267.4
GRPOModel=Qwen3-4B2026.0675.471614563.1
RLCSDModel=Olmo-3-7B2026.0675.476654866.1
OPSDModel=Qwen3-8B2026.0675.274614964.8
SRPOModel=Qwen3-4B2026.067571614563
SRPOModel=Qwen3-8B2026.0674.872604964
SDPOModel=Qwen3-4B2026.0674.472624563.4
GRPOModel=Olmo-3-7B2026.0673.869633961.2
RLSDModel=Olmo-3-7B2026.0673.865613859.4
OPSDModel=Qwen3-4B2026.0673.471624262.1
BaseModel=Qwen3-4B2026.0673.267584260.1
SDPOModel=Olmo-3-7B2026.0673.267594661.3
SRPOModel=Olmo-3-7B2026.0673.268614060.6
BaseModel=Qwen3-8B2026.0672.467554459.6
SDPOModel=Qwen3-8B2026.0672.472564661.6
OPSDModel=Olmo-3-7B2026.0672.469623860.4
BaseModel=Olmo-3-7B2026.0670.664553556.2
RLCSDModel=Qwen3-1.7B2026.067063633858.5
GRPOModel=Qwen3-1.7B2026.0667.459523453.1
SDPOModel=Qwen3-1.7B2026.0667.461543053.1
RLSDModel=Qwen3-1.7B2026.0666.859503552.7
OPSDModel=Qwen3-1.7B2026.0664.455523250.9
SRPOModel=Qwen3-1.7B2026.0664.458473350.6
BaseModel=Qwen3-1.7B2026.0663.253433147.6