BFCL
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
BFCL
36.6F1 Score
14
BFCL
46.9BFCL Score
14
BFCL
77.9Accuracy
14
BFCL v3
84.1Live Success Rate
14
BFCL Multi-Turn v3
69.1Overall Score
14
BFCL (Held-In)
89.4Accuracy
14
BFCL Non-Live
86.85Simple Accuracy
13
BFCL Non-Live Parallel Multiple
34BFCL Official Metric
12
BFCL Non-Live Parallel
63BFCL Official Score
12
BFCL Live Parallel Multiple
29.17BFCL Official Metric
12
BFCL Live Parallel
56.25BFCL Official Success Rate
12
BFCL
3.13Speedup
12
BFCL Average v4
63Accuracy
12
BFCL Non-Live v4
87.06Accuracy
12
BFCL Live v4
79.68Accuracy
12
BFCL Multiple
92.5Accuracy
12
BFCL Exec v3
94.6Overall Accuracy
12
BFCL Live v3
77.9Overall Accuracy
12
BFCL Non-Live v3
94Overall Accuracy
12
BFCL Opaque
78Semantic Similarity
12
BFCL V3
62.1Accuracy
12
BFCL Executable (test)
100Success Rate (Simple, Python)
12
BFCL
63.1Multi-turn Success Rate
11
BFCL V3 (test)
63.34Overall Accuracy
11
BFCL v4
72.4Base Accuracy
10