# vLLM v0.3.0 — Release & Upgrade Intelligence
- **Product:** vLLM (https://vllm.ai)
- **Version:** v0.3.0
- **Released:** Jan 31, 2024
- **Impact Score:** 5/10
- **Breaking Changes:** Not verified
- **Canonical Analysis:** https://changelogs.us/vllm/updates/v0.3.0/

## Executive Summary (TL;DR)
vLLM v0.3.0. Source highlight: Experimental multi-lora support

## New Features
- Add baichuan chat template jinjia file
- [DOC] Add additional comments for LLMEngine and AsyncLLMEngine
- Add gradio chatbot for openai webserver
- [CI] Add Buildkite
- Add StableLM3B model
- [Neuron] Add an option to build with neuron

## Improvements
- Optimized MoE performance and Deepseek MoE support

## Bug Fixes
- Miner fix of type hint
- Fix eager mode performance
- Fix ipv4 ipv6 dualstack
- [Minor] Fix the format in quick start guide related to Model Scope
- [BUG] RuntimeError: deque mutated during iteration in abort_seq_group
- Update quickstart.rst with small clarifying change (fix typo)

## Upgrade Command
```bash
python -m pip install 'vllm==0.3.0'
```