List your company

Tool Reasoning on ToolBench (G1)

85.5Pass RateToolGate
49.158.556877.45Jan 8, 2026

Evaluation results

MethodLinks
ToolGateModel=GPT-5.22026.0185.583.5
ToolGateModel=Gemini 3 Pro2026.018380.5
ToolChain*Model=GPT-5.22026.0182.880
LATSModel=GPT-5.22026.0180.378.8
ToolChain*Model=Gemini 3 Pro2026.018078.5
LATSModel=Gemini 3 Pro2026.0178.575.3
Tool-PlannerModel=GPT-5.22026.0175.572.3
Tool-PlannerModel=Gemini 3 Pro2026.0173.870
ToolGateModel=Deepseek V3.22026.017270.3
DFSDTModel=GPT-5.22026.017068.5
ToolChain*Model=Deepseek V3.22026.0168.865
ToolGateModel=Qwen-3-235B2026.0168.365.5
DFSDTModel=Gemini 3 Pro2026.0168.365.5
LATSModel=Deepseek V3.22026.0165.361.8
ToolChain*Model=Qwen-3-235B2026.016562.8
ReACTModel=GPT-5.22026.0163.562.8
LATSModel=Qwen-3-235B2026.0162.559.3
Tool-PlannerModel=Deepseek V3.22026.0162.560.3
Tool-PlannerModel=Qwen-3-235B2026.0160.358
ReACTModel=Gemini 3 Pro2026.016060.5
DFSDTModel=Deepseek V3.22026.0158.555
DFSDTModel=Qwen-3-235B2026.015753.8
ReACTModel=Deepseek V3.22026.015248.5
ReACTModel=Qwen-3-235B2026.0150.5—