List your company

Tool Use on RoTBench Multi-turn

72.9Tool Selection AccuracyPA-Tool
7.48424.46741.4558.433Oct 8, 2025

Evaluation results

MethodLinks
PA-ToolBackbone=Gemini-2.5-Flash, Model Category=Large Language Models2025.1072.964.3
BaseModel=GPT-4.1-mini2025.1071.461.4
GPT-4.1-miniMethod=Base, Model Category=Large Language Models2025.1071.461.4
PA-ToolBackbone=GPT-4.1-mini, Model Category=Large Language Models2025.107061.4
Gemini-2.5-Flash-LiteMethod=Base, Model Category=Small Language Models2025.1065.758.6
PA-ToolBackbone=Qwen3-4B, Model Category=Reasoning Models, Mode=thinking mode2025.1065.761.4
BaseModel=Claude-Sonnet-4.52025.1064.358.6
PA-ToolBackbone=Gemini-2.5-Flash-Lite, Model Category=Small Language Models2025.1064.358.6
PA-ToolBackbone=Llama3.3-70B, Model Category=Large Language Models2025.1064.348.6
BaseModel=Gemini-2.5-Flash2025.106054.3
Llama3.3-70BMethod=Base, Model Category=Large Language Models2025.106047.1
Gemini-2.5-FlashMethod=Base, Model Category=Large Language Models2025.106054.3
Qwen3-4BMethod=Base, Model Category=Reasoning Models, Mode=thinking mode2025.1057.152.9
Ministral-8BMethod=Base, Model Category=Small Language Models2025.1051.438.6
PA-ToolBackbone=Ministral-8B, Model Category=Small Language Models2025.1051.441.4
GPT-4.1-nanoMethod=Base, Model Category=Small Language Models2025.1051.444.3
PA-ToolModel=Llama3.1-8B2025.1048.635.7
PA-ToolBackbone=GPT-4.1-nano, Model Category=Small Language Models2025.1047.142.9
MostFreqModel=Llama3.1-8B2025.1045.734.3
PA-ToolBackbone=Qwen3-1.7B, Model Category=Reasoning Models, Mode=thinking mode2025.1045.738.6
GreedyModel=Llama3.1-8B2025.1044.332.9
BaseModel=Llama3.1-8B2025.1042.834.3
Qwen3-1.7BMethod=Base, Model Category=Reasoning Models, Mode=thinking mode2025.1038.638.6
PA-ToolModel=Llama3.2-3B2025.1034.328.6
BaseModel=Llama3.2-3B2025.1032.927.1
GreedyModel=Llama3.2-3B2025.1032.925.7
MostFreqModel=Llama3.2-3B2025.1032.927.1
PA-ToolModel=Qwen2.5-7B2025.1027.122.9
MostFreqModel=Qwen2.5-7B2025.1024.320
BaseModel=Qwen2.5-7B2025.1021.421.4
GreedyModel=Qwen2.5-7B2025.1018.615.7
PA-ToolModel=Qwen2.5-3B2025.1015.714.3
MostFreqModel=Qwen2.5-3B2025.1012.98.6
GreedyModel=Qwen2.5-3B2025.1011.47.1
BaseModel=Qwen2.5-3B2025.101010