List your company

Tool selection on API-Bench

96.3AccuracyMTA
32.204848.844965.48582.1251Jan 19, 2026

Evaluation results

MethodLinks
MTABackbone=Qwen1.5-Chat (7B)2026.0196.3
FTBackbone=Qwen1.5-Chat (7B)2026.0195.11
FTBackbone=Baichuan2-Chat (7B)2026.0194.77
MTABackbone=Chinese Alpaca-2 (7B)2026.0193.8
MTABackbone=Baichuan2-Chat (7B)2026.0193.3
FTBackbone=Chinese Alpaca-2 (7B)2026.0188.34
FTBackbone=ChatGLM3 (6B)2026.0183.12
MTABackbone=ChatGLM3 (6B)2026.0182.45
ReActBackbone=Qwen1.5-Chat (7B)2026.0170.33
ReActBackbone=ChatGLM3 (6B)2026.0165.34
CoTBackbone=Qwen1.5-Chat (7B)2026.0159.1
CoTBackbone=ChatGLM3 (6B)2026.0155.12
ReActBackbone=Baichuan2-Chat (7B)2026.0151.79
BaseBackbone=ChatGLM3 (6B)2026.0150.23
CoTBackbone=Baichuan2-Chat (7B)2026.0146.34
BaseBackbone=Qwen1.5-Chat (7B)2026.0145.02
BaseBackbone=Baichuan2-Chat (7B)2026.0142.48
ReActBackbone=Chinese Alpaca-2 (7B)2026.0139.12
CoTBackbone=Chinese Alpaca-2 (7B)2026.0138.91
BaseBackbone=Chinese Alpaca-2 (7B)2026.0134.67