Share your thoughts, 1 month free Claude Pro on usSee more

Agentic Task on AppWorld normal (test)

79.8Task Goal Completion (TGC)

SPO-GA

Updated 1mo ago

Evaluation Results

Method	Links
SPO-GA 2026.06		79.8
SAGE 2026.06		79.2
SPO-RS 2026.06		78
ACE offline 2026.06		76.2
DC-CU 2026.06		65.5
GEPA 2026.06		64.9
ICL 2026.06		64.3
Base LLM 2026.06		63.7