@QuixiAI: QuixiAI/ThunderMittens (fork from @HazyResearch) Porting ThunderKittens (and literally everything else) to Metal. Now w…
Summary
QuixiAI ported ThunderKittens to Metal, enabling kernel support on MPS and MLX for training models on Mac.
View Cached Full Text
Cached at: 06/30/26, 05:37 AM
QuixiAI/ThunderMittens (fork from @HazyResearch)
Porting ThunderKittens (and literally everything else) to Metal. Now works with MPS and MLX.
Why? I’ve a new model I’m cooking (inspired by @tri_dao and @_albertgu’s Mamba 3) and wanted to try training on my mac but, no kernels! And @TheEricHartford style, when I fix something, I fix it all the way down.
Similar Articles
@QuixiAI: QuixiAI/ThunderKittens and QuixiAI/ThunderMittens are now rebranded to QuixiCore-CUDA and QuixiCore-Metal Announcing Qu…
QuixiAI rebrands ThunderKittens and ThunderMittens into QuixiCore-CUDA and QuixiCore-Metal, creating a unified family of cross-platform kernels for AI workloads.
@QuixiAI: QuixiAI/ThunderMittens
QuixiAI announced ThunderMittens, a new AI model available on GitHub.
@no_stp_on_snek: wow look at the gains on metal!!!
The Qwen MLX Challenge is a competition for benchmarking AI models on Apple Silicon, with official scores and local iteration encouraged during validation.
@QuixiAI: https://x.com/QuixiAI/status/2073936537213915611
QuixiAI released QuixiCore, a family of native high-performance AI kernel libraries for modern accelerators, with standalone implementations for CUDA, Metal, ROCm, XPU, and Gaudi backends, all sharing a common contract but no shared code.
@zcbenz: MLX's implementation of RDMA (Remote Direct Memory Access) over Thunderbolt on macOS, can now be used as an independent…
MLX's RDMA-over-Thunderbolt implementation for macOS is now available as a standalone library, enabling high-speed Mac clusters for local AI workloads.