List your company

Tool-use on ToolBench (I1/I2/I3 Evaluation)

80.67Average Pass RateQwen2.5-7B-Instruct-CAST
3.845223.790143.73563.6799Dec 7, 2023May 3, 2024Sep 28, 2024Feb 23, 2025Jul 21, 2025Dec 16, 2025May 14, 2026

Evaluation results

MethodLinks
Qwen2.5-7B-Instruct-CASTCategory=Open Source2026.0580.67————————————79.43
Qwen2.5-Coder-7B-Instruct-CASTCategory=Open Source2026.0580.61————————————79.14
Llama-3.2-8B-Instruct-CASTCategory=Open Source2026.0575.93————————————47.23
Qwen2.5-7B-Instruct-GRPOCategory=Open Source2026.0572.71————————————68.23
Attention Buckets + DFSDT-RetrieverModel=ToolLlama, Parameters=7B2023.1271.368.5657065.5656784787164.5698971.5
Llama-3.2-8B-Instruct-GRPOCategory=Open Source2026.0571.23————————————45.23
Qwen2.5-Coder-7B-Instruct-GRPOCategory=Open Source2026.0571.13————————————67.82
DFSDTModel=GPT42023.1271.16067.571.567.86766.579.573.377.563.3718470.4
Qwen-MaxCategory=Closed Source2026.0570.93————————————71.83
DeepSeek-V3Category=Closed Source2026.0570.73————————————72.93
Llama-3.2-8B-Instruct-SFTCategory=Open Source2026.0570.63————————————45.83
Qwen2.5-7B-Instruct-SFTCategory=Open Source2026.0568.67————————————65.23
GraniteCategory=Other2026.0568.47————————————50.17
Attention Buckets + DFSDTModel=ToolLlama, Parameters=7B2023.126866.567.561.5626265.57871.57366.5678269.2
Qwen2.5-Coder-7B-Instruct-SFTCategory=Open Source2026.0567.84————————————65.37
DFSDT-RetrieverModel=ToolLlama2023.1267.36462.3645960.555.581.568.568.560.8657363.1
DFSDTModel=ToolLlama2023.1266.757556155.36254.57768.57758666960
OTCCategory=Other2026.0565.49————————————36.12
DFSDTModel=ChatGPT2023.1264.854.560.5656260.557.3757271.564.8626964.3
GPT-4oCategory=Closed Source2026.0564.43————————————67.63
GorillaCategory=Other2026.0562.27————————————46.29
SynWorldModel=GPT-4-turbo2025.0459.33————————————73
SynWorldModel=Qwen2-72B-Instruct2025.0458.52————————————73
ReACTModel=GPT42023.1257.253.5605058.853.563.56765.87260.3477864.4
SynWorldModel=Qwen-long2025.0457.2————————————81
Self-RefineModel=GPT-4-turbo2025.0456.8————————————73
DRAFTModel=Qwen2-72B-Instruct2025.0456.43————————————69
DRAFTModel=GPT-4-turbo2025.0454.83————————————72
Self-RefineModel=Qwen2-72B-Instruct2025.0454.33————————————65
DRAFTModel=Qwen-long2025.0454.2————————————79
Self-RefineModel=Qwen-long2025.0453.7————————————77
EasyToolModel=Qwen2-72B-Instruct2025.0452.97————————————58
PORToolBackbone=Qwen-2.5-7B-Instruct2025.1052.857.6—55.2—56.3—47.6—47.8—52.2——
EasyToolModel=GPT-4-turbo2025.0451.67————————————68
ARPOfmBackbone=Qwen-2.5-7B-Instruct2025.1050.958.4—49.6—56—47.9—50.3—43.4——
EasyToolModel=Qwen-long2025.0450.8————————————63
ReActModel=GPT-4-turbo2025.0450.67————————————67
ReActModel=Qwen2-72B-Instruct2025.0449.43————————————55
GRPOfmBackbone=Qwen-2.5-7B-Instruct2025.1049.455.1—50.7—57.2—38.1—46.7—48.4——
ToRLBackbone=Qwen-2.5-7B-Instruct2025.1049.352.7—49.1—54.5—41.7—48.5—49——
SimpleTIRBackbone=Qwen-2.5-7B-Instruct2025.1049.253.8—47.2—55—46.6—50.2—42.1——
ToolformerCategory=Other2026.0548.92————————————22.11
AgentFlowBackbone=Qwen-2.5-7B-Instruct2025.1048.552.6—45.3—54.2—38.4—50.8—49.5——
ReActModel=Qwen-long2025.0448.3————————————71
ToolAlignCategory=Other2026.0546.78————————————22.36
ReActCategory=Other2026.0543.37————————————18.22
DFSDTModel=Text-Davinci-0032023.1243.143.540.34443.84646.83740.54243.3466346.3
ReACTModel=ChatGPT2023.1240.241.5—44—44.5—42.5———22——
Attention Buckets + ReACTModel=ToolLlama, Parameters=7B2023.1231.331.5453242.533.54931.5653242285850.3
ReACTModel=ToolLlama2023.1229254529423347.530.550.831.541.8255547
DFSDTModel=Claude-22023.1222.620.5383144.318.543.31736.820.533.5286543.5
ReACTModel=Text-Davinci-0032023.1216.51228.52035.320318.529.814.529.8244533.2
ReACTModel=Claude-22023.126.85.5313.527.85.533.8635631.51447.534.4