●Change8

Migrating to vLLM v0.9.2rc2

Version v0.9.2rc2 introduces 1 breaking change. This guide details how to update your code.

Released: 7/6/2025

1
Breaking Changes
2
Migration Steps
8
Affected Symbols

⚠️ Check Your Code

If you use any of these symbols, you need to read this guide:

pplxBatchedTritonExpertsfused_marlin_moeClassifierPoolerCompressedTensorsW8A8Fp8MoECutlassMethodFlexAttentionllm.chatcompressed_tensors_moe

Breaking Changes

●Issue #1

Removed V0 backends for CPU, XPU, and TPU. Users must transition to the V1 engine for these hardware platforms.

Migration Steps

  1. 1
    Switch to V1 engine if using CPU, XPU, or TPU backends as V0 versions are removed.
  2. 2
    Update CI/build pipelines to include the new kvcache-connector dependency.

Release Summary

This release focuses on the transition to the V1 engine by removing legacy V0 backends for CPU/TPU/XPU, while adding support for Blackwell (SM100) and Llama 4. Key improvements include FP8 kernel optimizations, FlexAttention enhancements, and expanded multimodal support in the frontend.

Need More Details?

View the full release notes and all changes for vLLM v0.9.2rc2.

View Full Changelog