vLLM 0.30.0 deletes four things you may be running and shortens two model context windows without asking.

GPTQ group and dynamic activation ordering is gone: g_idx is now ignored, and the Marlin, GPTQ, CPU and RDNA3 kernels behind it were deleted. The scale-out endpoints /render, /derender and /inference/v1/generate no longer register on a plain vllm serve without --enable-scale-out, and the VLLM_ENABLE_SCALE_OUT_ENDPOINTS variable that used to turn them on was removed. Four items deprecated for 0.29 are now deleted outright, including the VLLM_PREFIX_CACHE_RETENTION_INTERVAL and VLLM_MM_HASHER_ALGORITHM environment variables and the ROCm CUDA_VISIBLE_DEVICES fallback.¹

Fast Start adds a persistent per-GPU weight-cache daemon that holds post-quantized, TP-sharded weights in GPU memory, so a restarting engine maps them over CUDA IPC with --load-format ipc_cache instead of reloading from disk. It now covers FP4 checkpoints and multi-node tensor parallelism.

Freezing the garbage collector during CUDA-graph capture cuts capture from 12s to 2s and engine init from 28.9s to 8.2s on an H200, per the release notes. Both figures are vLLM's own, measured on their hardware, not ours.

Also new: Gumbel-max watermarked generation with a keyed PRF and per-request opt-out, and HiSparse, a host-resident KV tier that spills sparse-MLA decode pages to pinned host memory under GPU pressure.