# vLLM v0.5.3 — Release & Upgrade Intelligence
- **Product:** vLLM (https://vllm.ai)
- **Version:** v0.5.3
- **Released:** Jul 23, 2024
- **Impact Score:** 5/10
- **Breaking Changes:** Not verified
- **Canonical Analysis:** https://changelogs.us/vllm/updates/v0.5.3/

## Executive Summary (TL;DR)
vLLM v0.5.3. Source highlight: vLLM now supports Meta Llama 3.1! Please checkout our blog here for initial details on running the model.

## New Features
- In order to support long context, a new rope extension method has been added and chunked prefill has been turned on by default for Meta Llama 3.1 series of model. (#6666, #6553, #6673)
- Add AWQ support to the Marlin kernel. This brings significant (1.5-2x) perf improvements to existing AWQ models! (#6612)
- [Docs] Add Google Cloud to sponsor list
- [Misc] Add CustomOp Interface to UnquantizedFusedMoEMethod
- [CI/Build][TPU] Add TPU CI test
- [ci][distributed] add pipeline parallel correctness test

## Improvements
- Many enhancements to TPU support. (#6277, #6457, #6506, #6504)
- Progress towards refactoring for SPMD worker execution. (#6032)
- Progress in improving prepare inputs procedure. (#6164, #6338, #6596)
- Memory optimization for pipeline parallelism. (#6455)
- The new `vllm` CLI is now ready for testing. It comes with three commands: `serve`, `complete`, and `chat`. Feedback and improvements are greatly welcomed! (#6431)
- [misc][distributed] improve tests

## Bug Fixes
- [BugFix][Model] Jamba - Handle aborted requests, Add tests and fix cleanup bug
- [Bugfix][CI/Build] Test prompt adapters in openai entrypoint tests
- [Misc] Fix typos in spec. decode metrics logging.
- [doc][misc] remind users to cancel debugging environment variables after debugging
- [Doc] Fix the lora adapter path in server startup script
- [ROCm] Cleanup Dockerfile and remove outdated patch

## Upgrade Command
```bash
python -m pip install 'vllm==0.5.3'
```