List your company

Task-goal completion on AppWorld normal (test)

75Average Task Completion ScoreLOOP
71.8872.6973.574.31Mar 12, 2026

Evaluation results

MethodLinks
LOOPAlgorithm=LOOP, Backbone=Qwen-2.5-32B2026.037578
DAPOAlgorithm=DAPO, Backbone=Qwen-2.5-32B2026.037477
ADAPOAlgorithm=ADAPO, Backbone=Qwen-2.5-32B2026.037378
RLOOAlgorithm=RLOO, Backbone=Qwen-2.5-32B2026.037278