Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Tool-use performance on Tau-bench Retail (test)

66Pass Rate

gpt-4o-FC-JOSH-SFT

14.904828.169941.43554.7001Sep 6, 2024
Updated 1mo ago

Evaluation Results

MethodLinks
2024.09
66
2024.09
65.21
2024.09
64.26
2024.09
63.13
2024.09
58.43
2024.09
58.26
2024.09
54.43
2024.09
50.78
2024.09
47.65
2024.09
44.6
2024.09
36.34
2024.09
16.87