TaskBench
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
TaskBench HuggingFace
77.1Function Selection Accuracy
45
TaskBench Multimedia
82.3Function Selection Acc
36
TaskBench DailyLife
96.8Function Selection Accuracy
36
TaskBench Daily Life
97.36Node-F1
25
TaskBench Multimedia
88.54Node F1
25
Taskbench-HF
0.75MRR
18
TaskBench Multimedia v1 (test)
88.54n-F1
14
TaskBench-MM
20.7F1 Score
12
TaskBench
21.1F1 Score
12
Taskbench DL
58.9Function Selection Accuracy
9
Taskbench-MM
27Function Selection Accuracy
9
TaskBench-23 2024 (N=501)
73.4ToolF1
8
TaskBench Multimedia APIs (test)
91.12Precision
1
TaskBench Hugging Face APIs (test)
74.97Precision
1
TaskBench Dailylife APIs (test)
95.01Precision
1