Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-turn tool calling on BFCL MT v4
Loading...
48
Success Rate
Pass@N (oracle)
13.68
22.59
31.5
40.41
Jun 24, 2026
Success Rate
Updated 1mo ago
Evaluation Results
Method
Method
Links
Success Rate
Pass@N (oracle)
Training=false, Backbo...
2026.06
48
Greedy Decoding
Training=false, Backbo...
2026.06
42.5
WildReward-8B
Training=true, Backbon...
2026.06
42.5
Progress Advantage
Training=false, Backbo...
2026.06
42.5
ThinkPRM-14B
Training=true, Backbon...
2026.06
40
Mean-of-N
Training=false, Backbo...
2026.06
38.5
ThinkPRM-7B
Training=true, Backbon...
2026.06
37
DeepConf Tail
Training=false, Backbo...
2026.06
36
Self-Certainty
Training=false, Backbo...
2026.06
34.5
DeepConf B10
Training=false, Backbo...
2026.06
34.5
Pass@N (oracle)
Training=false, Backbo...
2026.06
22
WildReward-8B
Training=true, Backbon...
2026.06
20
Greedy Decoding
Training=false, Backbo...
2026.06
19
ThinkPRM-14B
Training=true, Backbon...
2026.06
19
Progress Advantage
Training=false, Backbo...
2026.06
19
ThinkPRM-7B
Training=true, Backbon...
2026.06
18.5
Mean-of-N
Training=false, Backbo...
2026.06
17.5
DeepConf Tail
Training=false, Backbo...
2026.06
15.5
DeepConf B10
Training=false, Backbo...
2026.06
15.5
Self-Certainty
Training=false, Backbo...
2026.06
15
Feedback
Search any
task
Search any
task