News · Use it
vLLM 0.27.0
Kimi K3 full stack in one release, Qwen3.5 dense/MoE, PyTorch 2.13 (breaking env), deeper FlashAttention 4 on SM100, and a pile of DeepSeek-V4 serving wins. 561 commits, 242 contributors.10
Relevant to: self-hosters pinning a serving engine this week, especially Kimi K3 or DeepSeek-V4 traffic
What the verdict meansAvailable, works, worth your time today.
Sources
- github — vllm v0.27.0 · read
Originally part of the 11 August 2026 edition.