{"apiVersion":"v1","taxonomyVersion":1,"generatedAt":"2026-10-29T18:12:05.018Z","data":{"recordId":"wsa_3wpczijgyxrm","score":0.975,"confidence":"high","capped":false,"capReason":null,"sources":[{"n":1,"url":"https://github.com/vllm-project/vllm/releases/tag/v0.30.0","publisher":"GitHub watchlist","authority":"owner-controlled","verification":"verified","lastRetrievedAt":"2026-10-28T00:02:38.647Z","detectedPublishedAt":"2026-09-22T05:20:54.000Z","rights":"original"}],"claims":[{"id":"11661eb7-0d11-4b6d-8c2c-aa99dafd2bc4","kind":"availability","statement":"vLLM 0.30.0 was released on September 22.","citations":[{"source":1,"supportProbability":0.85,"supports":true,"quote":null}]},{"id":"8007d238-999a-4740-8f7f-e2206024d519","kind":"availability","statement":"The release provides Python wheels and Docker images for CUDA, ROCm, CPU, and XPU.","citations":[{"source":1,"supportProbability":0.88,"supports":true,"quote":null}]},{"id":"8ae22faa-783e-4042-b944-5ed1120100ef","kind":"capability","statement":"The release adds support for DeepSeek-V4.1-Flash, GLM-5.3-Flash, K2-Horizon, and other models.","citations":[{"source":1,"supportProbability":0.96,"supports":true,"quote":{"text":"GLM-5.3-Flash : EPLB support ( #55119 ), FlashKDA for KDA chunked prefill (1.7-3.8x faster than the Triton chunk path, #55737 ), dense and masked-MHA sparse prefill for the NoPE (256, 0, 256) layout ( #55738 ), decode hot-path cleanups removing per-layer copies and a duplicate…","rights":"original"}}]},{"id":"7a51421a-4e80-46f0-8ad5-12090ab22d26","kind":"capability","statement":"The release adds Gumbel-max watermarked generation and detection, with per-request opt-out.","citations":[{"source":1,"supportProbability":0.99,"supports":true,"quote":{"text":"Watermarking : Gumbel-max watermarked generation and detection with a keyed PRF, per-request opt-out and an example detection endpoint ( #54053 ); dual-key Gumbel-max makes it compatible with speculative decoding ( #56122 ); the Rust frontend forwards the per-request controls (…","rights":"original"}}]},{"id":"6939c1fb-9707-4671-bd84-d46c45957923","kind":"capability","statement":"HiSparse can spill sparse-MLA KV pages to pinned host memory under GPU pressure.","citations":[{"source":1,"supportProbability":0.98,"supports":true,"quote":{"text":"HiSparse : a host-resident tier for sparse-MLA decode that spills KV pages to pinned host memory under GPU pressure and serves top-k misses from a per-request GPU hot buffer, enabled through HiSparseConnector ( #53781 ), with Prometheus counters ( #56061 ), a host cache shared…","rights":"original"}}]},{"id":"96283a5d-595a-4233-b4c8-fd9fa850382d","kind":"capability","statement":"Model Runner V2 adds dual-batch overlap in eager mode and with FULL CUDA graphs.","citations":[{"source":1,"supportProbability":0.98,"supports":true,"quote":null}]}]},"nextCursor":null,"license":{"catalog":{"name":"CC BY 4.0","spdx":"CC-BY-4.0","url":"https://creativecommons.org/licenses/by/4.0/","attribution":"wowsuchai","attributionUrl":"https://wowsuchai.withglint.app","covers":["catalog metadata","classifications","generated summaries","feed event metadata"]},"excluded":{"covers":["source quotations","linked source content"],"terms":"Not licensed by wowsuchai. Source quotations and linked source content keep their original rights."}}}