@Kimi_Moonshot: We're open-sourcing FlashKDA — our high-performance CUTLASS-based implementation of Kimi Delta Attention kernels. Achie…
Summary
Moonshot AI releases FlashKDA, an open-source CUTLASS-based implementation of Kimi Delta Attention kernels that delivers 1.72×–2.22× prefill speedup on H20 GPUs.
Similar Articles
Moonshot open-sourced FlashKDA, CUTLASS kernels for Kimi Delta Attention, up to 2.22x over the Triton baseline on H20
MoonshotAI released FlashKDA, open-source CUTLASS kernels for Kimi Delta Attention that deliver up to 2.22x speedup over Triton on H20 GPUs.
MoonshotAI/FlashKDA
FlashKDA is a high-performance implementation of Kimi Delta Attention (KDA) kernels built on CUTLASS, optimized for SM90+ GPUs and integrated with flash-linear-attention.
@HotAisle: Kimi K2.6 + DFlash: 508 tok/s on 8x MI300X 5.6x throughput improvement over baseline autoregressive serving 90 tok/s → …
Kimi K2.6 paired with DFlash inference system achieves 508 tokens/s on 8×AMD MI300X, a 5.6× throughput jump from 90 tokens/s baseline with zero quality loss.
Kimi-K3 Technical Report [pdf]
MoonshotAI releases Kimi-K3, a 2.8T-parameter open-weight multimodal agentic model with a 1M-token context window, built on new Kimi Delta Attention and Attention Residuals architecture, achieving significant scaling improvements.
@akshay_pachaar: one matrix replaced the KV cache. (the technique is 100% open source) Kimi just dropped K3, an open model at frontier s…
Kimi released K3, a 2.8T-parameter open model using delta attention to avoid growing KV cache, enabling a 1-million-token context window with linear memory cost.