List your company

Tool-selection bias evaluation on Tool-selection bias benchmark

0.249δAPI BiasDeepSeek-V3.2-Exp
0.244160.276830.30950.34217Sep 30, 2025

Evaluation results

MethodLinks
DeepSeek-V3.2-Exp2025.090.2490.5040.377
ToolLLaMA2025.090.2770.3910.334
GPT-3.5-turbo2025.090.320.3360.328
Qwen3 235B2025.090.330.1680.249
GPT-4.1 mini2025.090.3310.4230.377
Gemini 2.5 Flash2025.090.3650.3060.335
Claude 3.5 Sonnet2025.090.370.3250.347