List your company

Travel Planning on TravelPlanner 1000 tasks (test)

94.72Commonsense Score (Micro)HiMAP-Travel
65.901673.383380.86588.3467Jan 30, 2026Feb 4, 2026Feb 10, 2026Feb 16, 2026Feb 21, 2026Feb 27, 2026Mar 5, 2026

Evaluation results

MethodLinks
HiMAP-TravelBackbone=Qwen3-4B2026.0394.72—66.0346.4463.4549.8
HiMAP-TravelBackbone=Qwen3-8B2026.0394.62—6750.4766.0552.65
DeepTravelBackbone=Qwen3-8B2026.0393.28—60.6245.6660.1243.98
AutoRefineAugmentation=ReAct2026.0191.7310042.242.137.234.1
MTPBackbone=GPT-4 w/ hint2026.0391.46—5053.9645.1242.68
DeepTravelBackbone=Qwen3-4B2026.0390.62—53.8539.552.3334.92
AutoRefine2026.0190.5999.937.937.1632.627.1
ATLASBackbone=Gemini-2.5-Pro2026.0385.81—40.577.6470.635
ReAct2026.0175.2699.216.549.0439.110.4
PMC2026.0173.8910015.5945.1933.5612.12
ATLAS2026.0173.8910015.5945.1933.5612.12
ReActAugmentation=Reflexion2026.0171.8499.813.6737.8426.79.13
ReActAugmentation=EvoAgent2026.0167.0198.891033.7120.426.11