BrowseComp
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
BrowseComp-ZH
9.69LJFT
5
BrowseComp Chinese version (test)
34.6Accuracy
5
BrowseComp-Plus
36.17Pass@1 Accuracy
5
BrowseComp (official)
71Exact Match
5
BrowseComp-Plus
64.7CC
4
BrowseComp
41.8CC
4
BrowseComp-Plus document corpus
74.7Recall
4
BrowseComp (test)
67.67Accuracy
4
BrowseComp (dev)
72.5Accuracy
4
BrowseComp
16.7Task Completion Score (TCS)
4
BrowseComp w/ Context Manage
67.6Score
4
BrowseComp-zh
82Score
3
BrowseComp-Plus
51.45Accuracy
3
BrowseComp
19.7Pass@3
3
BrowseComp
34Pass@3
3
BrowseComp-Plus
27.1Accuracy
3
BrowseComp lite
45.3Accuracy
3
BrowseComp EN
32Accuracy
2
BrowseComp
73.3Score
2
BrowseComp with Search
33.89Success Rate
2