vLLM v0.28.0 merged 584 commits from 270 contributors, with end-to-end DeepSeek V4 sparse MLA.

The August 26 release adds Decode Context Parallel for Kimi-K3, fused FlashKDA decode and prefill kernels, combined all-gathers with a claimed 1.5-3x kernel-level speedup, an adaptive speculative token budget worth about 60% better DSpark TTFT, and optional shared-expert sharding saving about 17 GiB per GPU (GitHub). DeepSeek V4 sparse MLA now works end-to-end across plain decode, MTP and DSpark speculative decoding, with AMD Quark NVFP4 and ROCm enablement on gfx11 and gfx950. If you self-host frontier open weights, this is the release that makes the two newest Chinese flagships cheap to serve.