List your company

Physics-based puzzle solving on INTERPHYRE 1.0 (test)

100Cliffhanger ScoreClaude Sonnet
-4235077Jun 28, 2026

Evaluation results

MethodLinks
Claude SonnetBackbone=Claude Sonnet, Evaluation Protocol=REACT baseline2026.06100—96100100100100248
O-120BBackbone=O-120B, Evaluation Protocol=REACT baseline2026.0664—90.13372.577.8100220
O-20BBackbone=O-20B, Evaluation Protocol=REACT baseline2026.0630.5—8014352658160
Q-14BBackbone=Q-14B, Evaluation Protocol=REACT baseline2026.066—42107224529.80
Q-7BBackbone=Q-7B, Evaluation Protocol=REACT baseline2026.062—2486218207.80
Random AgentMethodology=Heuristic policy, Sampling=Uniformly from valid placement region2026.060—34814100
Qwen-2.5-7B InstructProtocol=DIRECT baseline, Max Interaction Turns=252026.060—00184800
Qwen-2.5-14B InstructProtocol=DIRECT baseline, Max Interaction Turns=252026.060—32016100
GPT-OSS-20BProtocol=DIRECT baseline, Max Interaction Turns=252026.060—8141613300
GPT-OSS-120BProtocol=DIRECT baseline, Max Interaction Turns=252026.060—213312930
Claude SonnetProtocol=DIRECT baseline, Max Interaction Turns=252026.060—0914682302
RandBackbone=Random, Evaluation Protocol=REACT baseline2026.060—34814100