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

VerifyBench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Verifiable JudgingVerifyBench Hard
Accuracy88.8
19
Verifiable JudgingVerifyBench (full)
Accuracy95.7
19
VerificationVerifyBench Hard 1.0 (test)
Mean@3 Accuracy91.9
18
VerificationVerifyBench 1.0 (test)
m@3 Accuracy96.6
18
Showing 4 of 4 rows