BFCL
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
BFCL Multiple v3
84.5Accuracy
8
BFCL Agentic v4 (out-of-distribution)
39Web-base Score
8
BFCL v2
1Spearman Rho (FullPair Agreement)
7
BFCL BASE and LONG CONTEXT multi-turn v3
41.88Avg@4 Success Rate
7
BFCL (train)
75.04Success Rate (SR)
7
BFCL (test)
62.38Success Rate
7
BFCL Live
86.8Success Rate (Simple)
7
BFCL Non-Live
76.1Simple Success Rate
7
BFCL
69.9Live Success Rate
7
BFCL V3
70.4pass@1
7
BFCL Multi-turn V4
58.5Base Score
7
BFCL Simple (Live)
80.1Accuracy
6
BFCL Simple split (Non-live)
88.6Accuracy
6
BFCL V4
18Average Performance
6
BFCL live_multiple
94.3Free Name Accuracy
6
BFCL v4 (non-live and live)
86.45Accuracy
6
BFCL Multi-turn V4
60Base Success Rate
6
BFCL When2Call v2
80ToolCall Precision
6
BFCL
70Accuracy
6
BFCL Multiple v1 (test)
92Accuracy
6
BFCL Simple Python v1 (test)
93.5Accuracy
6
BFCL V4
53.8avg@3
6
BFCL v4
16Web Search Accuracy
6
BFCL v3 (test)
18Base Accuracy
6
BFCL v4
68.14Accuracy
5