KernelBenchMega
KernelBenchCUDA
GLM-5.2 Fused MoEDeepSeek NSAMegaQwen DecodeGrid + MinGRU SPS
GLM-5.2 Fused MoE
DeepSeek NSA
MegaQwen Decode
KernelBenchHard
FP8 GEMMKDA CUTLASSPaged AttentionTopK BitonicSonic MoEW4A16 GEMM
FP8 GEMM
KDA CUTLASS
Paged Attention
TopK Bitonic
Sonic MoE
KernelBenchMultisoon
pass bar = share of best · click a cell for solution / traceblank no run yetwrong answers don't matchbuild can't compile/importslow timed outcut stopped earlyflag audit reject