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

APT-Bench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Atomic Perception of TransitionsAPT-BENCH (test)
Recall53.4
12
APT chain enumerationAPT-BENCH (test)
Recall14
8
ToolAPT-Bench
Accuracy65.8
6
MathAPT-Bench
Accuracy70.5
6
Deep ResearchAPT-Bench
Accuracy40.5
6
CodeAPT-Bench
Accuracy41.9
6
Showing 6 of 6 rows