●Change8

Migrating to vLLM v0.9.2

Version v0.9.2 introduces 3 breaking changes. This guide details how to update your code.

Released: 7/7/2025

3
Breaking Changes
4
Migration Steps
9
Affected Symbols

⚠️ Check Your Code

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

LLM.beam_searchllm.chatFusedMoEModularKernelAutoWeightsLoaderMultiModalHasher.hash_prompt_mm_dataCachedRequestDataFlashAttentionFlashMLATritonAttention

Breaking Changes

●Issue #1

FP4 emulation has been removed on devices with compute capability less than SM100 (Blackwell).

●Issue #2

Runtime (cloud)pickle imports are now forbidden for security hardening.

●Issue #3

This is the final version where V0 engine code and features remain intact; future versions will prioritize V1.

Migration Steps

  1. 1
    Migrate from V0 engine to V1 engine as V0 will be removed/changed in the next version.
  2. 2
    Update monitoring systems to use new metric names as 'gpu_' prefixes are deprecated for general metrics.
  3. 3
    Ensure hardware compatibility for FP4 features (requires SM100+).
  4. 4
    Review custom code for any reliance on cloudpickle imports, which are now restricted.

Release Summary

This release marks the final transition phase to the V1 engine, introducing Blackwell (SM100/120) support, Expert-Parallel Load Balancing, and expanded multi-modal/audio API capabilities. It includes significant performance optimizations for CUDA-Graphs and broadens hardware support for Intel GPUs and TPUs.

Need More Details?

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

View Full Changelog