# vLLM v0.20.2 — Release & Upgrade Intelligence
- **Product:** vLLM (https://vllm.ai)
- **Version:** v0.20.2
- **Released:** May 10, 2026
- **Impact Score:** 5/10
- **Breaking Changes:** Not verified
- **Canonical Analysis:** https://changelogs.us/vllm/updates/v0.20.2/

## Executive Summary (TL;DR)
vLLM v0.20.2. Source highlight: This release features 6 commits from 6 contributors (0 new)!

## Bug Fixes
- **DeepSeek V4 sparse attention**: Re-enable the persistent topk path on Hopper and ensure the memset kernel runs at CUDA graph capture time regardless of `max_seq_len`, fixing the MTP=1 hang on DeepSeek V4 (#41665, revert of #41605).
- **DeepSeek V4 KV cache**: Fixed a "failure to allocate KV blocks" error in the V1 engine KV cache manager (#41282).
- **gpt-oss MXFP4 + torch.compile**: Plumbed `hidden_dim_unpadded` through the `moe_forward` fake op so MXFP4 works under `torch.compile` on v0.20.x (#42002, backport of #41646).
- **Qwen3-VL**: Removed an invalid deepstack boundary check that could fail under heavy load (#40932).

## Upgrade Command
```bash
python -m pip install 'vllm==0.20.2'
```