v0.20.2

mailcow/mailcow-dockerizedv0.20.2May 10, 2026by khluu

AI Summary

This is a small patch release that addresses bug fixes for DeepSeek V4, gpt-oss, and Qwen3-VL models. It resolves stability issues related to attention mechanisms, KV cache allocation, and compatibility with PyTorch compilation.

Key Highlights

Fixed DeepSeek V4 sparse attention hang on Hopper architecture.
Resolved KV cache allocation failures in DeepSeek V4 engine.
Enabled MXFP4 support under torch.compile for gpt-oss.
Fixed boundary checks in Qwen3-VL to prevent failures under heavy load.

Full Release Notes

# vLLM v0.20.2

## Highlights
This release features 6 commits from 6 contributors (0 new)!

This is a small patch release with bug fixes for DeepSeek V4, gpt-oss, and Qwen3-VL

### 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).

## Contributors
@ywang96, @zyongye, @stecasta, @wzhao18, @Isotr0py, @khluu