Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

RepoEval

Benchmarks

Task NameDataset NameSOTA ResultTrend
API Invocation CompletionRepoEval 1.0 (test)
Exact Match50.13
24
Line CompletionRepoEval 1.0 (test)
Exact Match57.75
24
Repository-level code-completionRepoEval (test)
Exact Match49.9
8
Function Body CompletionRepoEval Function Body Completion (All)
Pass Rate42.63
6
Function CompletionRepoEval leopard (ai-betty)
Pass@564.99
3
Function CompletionRepoEval deepmind tracr
Pass@556.11
3
Function CompletionRepoEval amazon-science patchcore-inspection
Pass@566.94
3
Showing 7 of 7 rows