List your company

Agent Task on AppWorld Normal (test)

92.3TGCACCORD
-2.454422.145346.74571.3447Oct 6, 2025Nov 17, 2025Dec 29, 2025Feb 9, 2026Mar 23, 2026May 4, 2026Jun 15, 2026

Evaluation results

MethodLinks
ACCORDBackbone=Claude-4.5-sonnet2026.0692.387.5
ReActBackbone=Claude-4.5-sonnet2026.0688.183.9
ACCORDBackbone=Qwen3.5-27B-FP82026.0681.567.9
ACCORDBackbone=GPT-5-mini2026.067862.5
ReAct + ACEBase LLM=DeepSeek-V3.1-671B, Adaptation Mode=Offline Adaptation, GT Labels=✓2025.1076.264.3
ReAct + ACEBase LLM=DeepSeek-V3.1-671B, Adaptation Mode=Offline Adaptation, GT Labels=✗2025.107564.3
Claude-Sonnet-4.5CoEvolve=false2026.0473.8155.36
ReActBackbone=Qwen3.5-27B-FP82026.0671.451.8
ReAct + ACEBase LLM=DeepSeek-V3.1-671B, Adaptation Mode=Online Adaptation, GT Labels=✗2025.1069.653.6
ReAct + DC (CU)Base LLM=DeepSeek-V3.1-671B, Adaptation Mode=Online Adaptation, GT Labels=✗2025.1065.558.9
ReAct + GEPABase LLM=DeepSeek-V3.1-671B, Adaptation Mode=Offline Adaptation, GT Labels=✓2025.1064.944.6
ReAct + ICLBase LLM=DeepSeek-V3.1-671B, Adaptation Mode=Offline Adaptation, GT Labels=✓2025.1064.346.4
ReActBase LLM=DeepSeek-V3.1-671B, Adaptation Mode=None2025.1063.742.9
ReActBackbone=GPT-5-mini2026.0663.746.4
Self-RefineBackbone=GPT-5-mini2026.0660.139.3
Qwen3-30B-A3BCoEvolve=true2026.0454.7633.93
Gemini-2.5-FlashCoEvolve=false2026.0453.5732.14
DeepSeek-V3.2CoEvolve=false2026.0447.0230.36
FullCodeReflexBackbone=GPT-5-mini2026.0643.530.4
Qwen3-4B-InstructCoEvolve=true2026.0435.7114.28
Qwen3-30B-A3BCoEvolve=false2026.0431.5512.5
LLaMA-3.3-70BCoEvolve=false2026.0431.548.92
GPT-4CoEvolve=false2026.0430.421.4
Qwen2.5-7B-InstructCoEvolve=true2026.0427.9812.5
Gemma-3-27BCoEvolve=false2026.0423.817.14
ACE (online)Method Category=Cross-trajectory memory methods2026.0621.810.2
GPT-OSS-20BCoEvolve=false2026.0417.863.57
Qwen3-4B-InstructCoEvolve=false2026.0416.675.36
Qwen2.5-7B-InstructCoEvolve=false2026.041.190