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

LIVECODE

Benchmarks

Task NameDataset NameSOTA ResultTrend
CodingLiveCode (test)
Score36.8
32
CodingLiveCode
LiveCode Score41.2
31
CodeLiveCode
Accuracy83.6
30
Code generationLiveCode
Acceptance Length (τ)2.975
16
Code GenerationLiveCode
Pass@142.8
15
Code GenerationLIVECODE (held-out)
DBS25.7
9
Code GenerationLiveCode (test)
Accuracy67.8
8
Showing 7 of 7 rows