List your company

Agent Tool-Use Planning on ThinkGeo 14 Tools (test)

77.7Tool Selection Accuracy (TSA)Qwen3-235B
47.12455.0626370.938May 6, 2026

Evaluation results

MethodLinks
Qwen3-235BApproach=Meta2026.0577.774.35530.788.676.762.5
Qwen3-8BApproach=Meta2026.0574.374.151.924.676.965.950.6
Qwen2.5-72BApproach=Meta2026.05727249.721.972.762.539.5
GPT-5Approach=Meta2026.0552.952.825.94084.173.150.9
Qwen2.5-72BApproach=Base2026.0551.151.134.426.371.859.28
Qwen3-8BApproach=Base2026.0550.250.134.819.976.960.614.8
GPT-5Approach=Base2026.0548.448.429.353.570.957.139.1
Qwen3-235BApproach=Base2026.0548.345.539.22984.569.134.8