vLLM Releases Version 0.25.0
vllm_project · x · 2026-07-12
vLLM officially releases v0.25.0, a massive update comprising 558 commits and 232 contributors.
Key highlights:
- Model Runner V2 becomes the default execution path for all dense models.
- Legacy PagedAttention is retired.
- Transformers backend speed is now close to native vLLM.
- Introduced a unified Streaming Parser Engine.
- Added universal speculative decoding (TLI) across heterogeneous vocabularies, alongside new DSpark and DFlash drafters.
- New model support including Hy3 and Unlimited OCR.
This is the main release announcement for the vLLM upgrade, which subsequent detailed serving and hardware notes are based upon.
More from Infra
- Anthropic S-1 reveals $800B+ in off-books future compute commitments — GaryMarcus · 2026-10-27
- ~42% of Anthropic's ~11GW contracted compute is NVIDIA GPUs — ns123abc · 2026-10-27
- EdgeRock-4B holds KV cache at 72 MiB across 256K context, 99.1% smaller than Qwen3.5-4B — alifcoder · 2026-10-27
- RTX 5080 price jumps from $1658 to $2318 in ten days on Amazon — Frizzy-MacDrizzle · 2026-10-27
- ComfyUI on 12GB VRAM: measured peaks show offloading turns 24s into 11 minutes — Impossible-Mail-4619 · 2026-10-27
- Neon rebuilds its docs search entirely on its own stack: Lakebase Search, Functions, AI Gateway — matei_zaharia · 2026-10-27