BFCL
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
BFCL v2 (test)
67.2Recall@1
4
BFCL (out domain)
69Success Rate
4
BFCL MT
33.5Score
4
BFCL live and Irrelevance Detection 2025
88BFCL live Score
4
BFCL live and Irrelevance Detection
86.2BFCL live Score
4
BFCL Multi Turn
59.4Task Completion Score (TCS)
4
BFCL Web Search
86.5Task Completion Score (TCS)
4
BFCL multi-turn category (test)
34.48Accuracy
4
BFCL unseen (test)
48.5Accuracy
4
BFCL rand (test)
56.6Accuracy
4
BFCL non-live (test)
82AST Accuracy (Simple Python)
4
BFCL (test)
88Recall@1
3
BFCL queries requiring merged tools
16Number Correct
3
BFCL Generalization (test)
87.5Final Accuracy
3
BFCL MULTI_TURN_LIVE
0.195Accuracy
3
BFCL AST LIVE
80.61Accuracy
3
BFCL AST NON LIVE
87.36Accuracy
3
BFCL Parity
86.2Accuracy
2
BFCL
73.69Score
2
BFCL v3 (val)
54.2Base Score
2
BFCL (Berkeley Function Calling Leaderboard) zero-shot transfer
87.6Agreement
2
BFCL live 2025
55.43Simple Success Rate
2
BFCL live 2025
46.68Simple Retrieval Success Rate
2
BFCL Non-live AST
96.5Success Rate (Python)
2
BFCL V4
24.52Memory Sum Component
2