List your company

Tool Use on ToolBench Hard v1.0

97Tool Call Rate (TCR)NaviAgent
21.91241.40660.980.394Jun 24, 2025

Evaluation results

MethodLinks
NaviAgentModel=Deepseek-V32025.069744.95.19
NaviAgentModel=Qwen2.5-32B2025.0679.430.85.31
α-UMIModel=Qwen2.5-32B2025.0677.122.96.58
α-UMIModel=Deepseek-V32025.067329.56.64
α-UMIModel=Qwen2.5-14B2025.0667.613.36.07
NaviAgentModel=Qwen2.5-14B2025.0661.122.44.68
ToolLLMModel=Qwen2.5-32B2025.0649.523.84.1
ReActModel=Deepseek-V32025.0648.623.83.74
ToolLLMModel=Qwen2.5-14B2025.0638.111.44.06
ReActModel=Qwen2.5-32B2025.0630.5193.95
ToolLLMModel=Deepseek-V32025.0629.724.83.9
ReActModel=Qwen2.5-14B2025.0624.8203.64