kernel-bench

Tag

Cards List
#kernel-bench

@elliotarledge: For those wondering why I use a Kimi Linear megakernel instead of Qwen 3.6, first look at the parameter counts. One is …

X AI KOLs Timeline · 2026-07-03 Cached

Elliot Arledge explains why he prefers using a Kimi Linear megakernel over Qwen 3.6 for kernel performance, comparing parameter counts, layer synchronization, hidden dimensions, and architecture-specific optimizations. The discussion highlights that Kimi Linear architecture is more suitable for megakernel implementation, especially for batch-1 decode on RTX PRO 6000 Blackwell.

0 favorites 0 likes
#kernel-bench

@VukRosic99: MusaCoder trained a 27B model to write native CUDA/MUSA kernels - and reached 93.2 Pass@8 on KernelBench vs Claude Opus…

X AI KOLs Timeline · 2026-06-27 Cached

MusaCoder trained a 27B model that achieves 93.2 Pass@8 on KernelBench, outperforming Claude Opus (87.2), and a 5-page technical summary of its data pipeline, verifier, and RL stabilizers is provided.

1 favorites 1 likes
← Back to home

Submit Feedback