# vLLM v0.6.6 — Release & Upgrade Intelligence
- **Product:** vLLM (https://vllm.ai)
- **Version:** v0.6.6
- **Released:** Dec 27, 2024
- **Impact Score:** 7/10
- **Breaking Changes:** Source flagged
- **Canonical Analysis:** https://changelogs.us/vllm/updates/v0.6.6/

## Executive Summary (TL;DR)
vLLM v0.6.6. Source highlight: Support Deepseek V3 (#11523, #11502) model.

## Breaking Changes & Deprecations
- Breaking change: `X-Request-ID` echoing is now opt-in instead of on by default for performance reason. Set `--enable-request-id-headers` to enable it.
- [V1] [1/N] [Breaking Change] API Server  (Remove Proxy)  by @robertgshaw2-neuralmagic in https://github.com/vllm-project/vllm/pull/11529

## New Features
- Add `QVQ` and `QwQ` to the list of supported models (#11509)
- Add pypi index for every commit and nightly build (#11404)
- [misc] benchmark_throughput : Add LoRA
- [Feature] Add load generation config from model
- [Model] Add JambaForSequenceClassification model
- [Misc] Add tqdm progress bar during graph capture

## Improvements
- We are just getting started for enhancing the support and unlock more performance. See #11539 for planned work.
- Breaking change: `X-Request-ID` echoing is now opt-in instead of on by default for performance reason. Set `--enable-request-id-headers` to enable it.
- Cutlass 2:4 Sparsity + FP8/INT8 Quant Support (#10995)
- [misc][perf] remove old code
- [Doc] Improve GitHub links

## Bug Fixes
- [Bugfix] Set temperature=0.7 in test_guided_choice_chat
- [Bugfix] Restore support for larger block sizes
- [Bugfix] Fix guided decoding with tokenizer mode mistral
- [MISC][XPU]update ipex link for CI fix
- [Bugfix] Fix broken phi3-v mm_processor_kwargs tests
- [FIX] update openai version

## Upgrade Command
```bash
python -m pip install 'vllm==0.6.6'
```