v0.14.0

## Highlights
This release features approximately 180 commits from over 70 contributors (23 new contributors).

vLLM-Omni **v0.14.0** is a feature-heavy release that expands Omni’s **diffusion / image-video generation** and **audio / TTS** stack, improves **distributed execution and memory efficiency**, and broadens **platform/backend coverage** (GPU/ROCm/NPU/XPU). It also brings meaningful upgrades to **serving APIs**, **profiling & benchmarking**, and overall **stability**.

### Key Improvements:

* **Async chunk** ([#727]): chunk pipeline overlap across stages to reduce idle time and improve end-to-end throughput/latency for staged execution.
* **Stage-based deployment for the Bagel model** ([#726]): Multi-stage pipeline (Thinker/AR stage + Diffusion/DiT stage) aligning it with the vllm-omni architecture
* **Qwen3-TTS model family support** ([#895]): Expands text-to-audio generation and supports online serving.
* **Diffusion LoRA Adapter Support (PEFT-compatible)** ([#758]): Adds LoRA fine-tuning/adaptation for diffusion workflows with a PEFT-aligned interface.
* **DiT layerwise (blockwise) CPU offloading** ([#858]): Fine-grained offloading to increase memory headroom for larger diffusion runs.
* **Hardware platforms + plugin system** ([#774]): Establishes a more extensible platform capability layer for cleaner multi-backend development.

---

### Diffusion & Image/Video Generation

* **Sequence Parallelism (SP) foundations + expansion**: Adds a non-intrusive SP abstraction for diffusion models ([#779]), SP support in LongCatImageTransformer ([#721]), and SP support for Wan2.2 diffusion ([#966]).
* **CFG improvements and parallelization**: CFG parallel support for Qwen-Image ([#444]), CFG parallel abstraction ([#851]), and online-serving CFG parameter support ([#824]).
* **Acceleration & execution plumbing**: Torch compile support for diffusion ([#684]), GPU diffusion runner ([#822]), and diffusion executor ([#865]).
* **Caching and memory efficiency**: TeaCache for Z-Image ([#817]) and TeaCache for Bagel ([#848]); plus CPU offloading for diffusion ([#497]) and DiT tensor parallel enablement for diffusion pipeline (Z-Image) ([#735]).
* **Model coverage expansion**: Adds GLM-Image support ([#847]), FLUX family additions (e.g., FLUX.1-dev [#853], FLUX.2-klein [#809]) and related TP support ([#973]).
* **Quality/stability fixes for pipelines**: Multiple diffusion pipeline correctness fixes (e.g., CFG parsing failure fix [#922], SD3 compatibility fix [#772], video saving bug under certain fps [#893], noisy output without a seed in Qwen Image [#1043]).

---

### Audio & Speech (TTS / Text-to-Audio)

* **Text-to-audio model support**: Stable Audio Open support for text-to-audio generation ([#331]).
* **Qwen3-TTS stack maturation**: Model series support ([#895]), online serving support ([#968]), plus stabilization fixes such as profile-run hang resolution ([#1082]) and dependency additions for Qwen3-TTS support ([#981]).
* **Interoperability & correctness**: Fixes and improvements across audio outputs and model input validation (e.g., StableAudio output standardization [#842], speaker/voices loading from config [#1079]).

---

### Serving, APIs, and Frontend

* **Diffusion-mode service endpoints & compatibility**: Adds /health and /v1/models endpoints for diffusion mode and fixes streaming compatibility ([#454]).
* **New/expanded image APIs**: /v1/images/edit interface ([#1101]).
* **Online serving usability improvements**: Enables tensor_parallel_size argument with online serving command ([#761]) and supports CFG parameters in online serving ([#824]).
* **Batching & request handling**: Frontend/model support for batch requests (OmniDiffusionReq refinement) ([#797]).

---

### Performance & Efficiency

* **Qwen3-Omni performance work**: SharedFusedMoE integration ([#560]), fused QKV & projection optimizations (e.g., fuse QKV linear and gate_up proj [#734], Talker MTP optimization [#1005]).
* **Attention and kernel/backend tuning**: Flash Attention attention-mask support ([#760]), FA3 backend defaults when supported ([#783]), and ROCm performance additions like AITER Flash Attention ([#941]).
* **Memory-aware optimizations**: Conditional transformer loading for Wan2.2 to reduce memory usage ([#980]).

---

### Hardware / Backends / CI Coverage

* **Broader backend support**: XPU backend support ([#191]) plus the platform/plugin system groundwork ([#774]).
* **NPU & ROCm updates**: NPU upgrade alignment ([#820], [#1114]) and ROCm CI expansion / optimization ([#542], [#885], [#1039]).
* **Test reliability / coverage**: CI split to avoid timeouts ([#883]) and additional end-to-end / precision tests (e.g., chunk e2e tests [#956]).

---

### Reliability, Correctness, and Developer Experience

* **Stability fixes across staged execution and serving**: Fixes for stage config loading issues ([#860]), stage output mismatch in online batching ([#691]), and server readiness wait-time increase for slow model loads ([#1089]).
* **Profiling & benchmarking improvements**: Diffusion profiler support ([#709]) plus benchmark additions (e.g., online benchmark [#780]).
* **Documentation refresh**: Multiple diffusion docs refactors and new guides (e.g., profiling guide [#738], torch profiler guide [#570], diffusion docs refactor [#753], ROCm instructions updates [#678], [#905]).

---

## What's Changed
* [Docs] Fix diffusion module design doc by @SamitHuang in https://github.com/vllm-project/vllm-omni/pull/645
* [Docs] Remove multi-request streaming design document and update ray-based execution documentation structure by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/641
* [Bugfix] Fix TI2V-5B weight loading by loading transformer config from model by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/633
* Support sleep, wake_up and load_weights for Omni Diffusion by @knlnguyen1802 in https://github.com/vllm-project/vllm-omni/pull/376
* [Misc] Merge diffusion forward context by @iwzbi in https://github.com/vllm-project/vllm-omni/pull/582
* [Doc] User guide for torch profiler  by @lishunyang12 in https://github.com/vllm-project/vllm-omni/pull/570
* [Docs][NPU] Upgrade to v0.12.0 by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/656
* [BugFix] token2wav code out of range by @Bounty-hunter in https://github.com/vllm-project/vllm-omni/pull/655
* [Doc] Update version 0.12.0 by @ywang96 in https://github.com/vllm-project/vllm-omni/pull/662
* [Docs] Update diffusion_acceleration.md by @SamitHuang in https://github.com/vllm-project/vllm-omni/pull/659
* [Docs] Guide for using sleep mode and enable sleep mode by @knlnguyen1802 in https://github.com/vllm-project/vllm-omni/pull/660
* [Diffusion][Feature] CFG parallel support for Qwen-Image by @wtomin in https://github.com/vllm-project/vllm-omni/pull/444
* [BUGFIX] Delete the CUDA context in the stage process. by @fake0fan in https://github.com/vllm-project/vllm-omni/pull/661
* [Misc] Fix docs display problem of streaming mode and other related issues by @Gaohan123 in https://github.com/vllm-project/vllm-omni/pull/667
* [Model] Add Stable Audio Open support for text-to-audio generation by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/331
* [Doc] Update ROCm getting started instruction by @tjtanaa in https://github.com/vllm-project/vllm-omni/pull/678
* [Bugfix] Fix f-string formatting in image generation pipelines by @ApsarasX in https://github.com/vllm-project/vllm-omni/pull/689
* [Bugfix] Solve Ulysses-SP sequence length not divisible by SP degree (using padding and attention mask) by @wtomin in https://github.com/vllm-project/vllm-omni/pull/672
* omni entrypoint support tokenizer arg by @divyanshsinghvi in https://github.com/vllm-project/vllm-omni/pull/572
* [Bug fix] fix e2e_total_tokens and e2e_total_time_ms by @LJH-LBJ in https://github.com/vllm-project/vllm-omni/pull/648
* [BugFix] Explicitly release file locks during stage worker init by @yuanheng-zhao in https://github.com/vllm-project/vllm-omni/pull/703
* [BugFix] Fix stage engine outputs mismatch bug in online batching by @ZeldaHuang in https://github.com/vllm-project/vllm-omni/pull/691
* [core] add torch compile for diffusion by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/684
* [BugFix] Remove duplicate width assignment in SD3 pipeline by @dongbo910220 in https://github.com/vllm-project/vllm-omni/pull/708
* [Feature] Support Qwen3 Omni talker cudagraph by @ZeldaHuang in https://github.com/vllm-project/vllm-omni/pull/669
* [Benchmark] DiT Model Benchmark under Mixed Workloads by @asukaqaq-s in https://github.com/vllm-project/vllm-omni/pull/529
* update design doc by @hsliuustc0106 in https://github.com/vllm-project/vllm-omni/pull/711
* [Perf] Use vLLM's SharedFusedMoE in Qwen3-Omni by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/560
* [Doc]: update vllm serve param and base64 data truncation by @nuclearwu in https://github.com/vllm-project/vllm-omni/pull/718
* [BugFix] Fix assuming all stage model have talker by @princepride in https://github.com/vllm-project/vllm-omni/pull/730
* [Perf][Qwen3-Omni] Fuse QKV linear and gate_up proj by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/734
* [Feat] Enable DiT tensor parallel for Diffusion Pipeline(Z-Image) by @dongbo910220 in https://github.com/vllm-project/vllm-omni/pull/735
* [Bugfix] Fix multi-audio input shape alignment for Qwen3-Omni Thinker by @LJH-LBJ in https://github.com/vllm-project/vllm-omni/pull/697
* [ROCm] [CI] Add More Tests by @tjtanaa in https://github.com/vllm-project/vllm-omni/pull/542
* [Docs] update design doc templated in RFC by @hsliuustc0106 in https://github.com/vllm-project/vllm-omni/pull/746
* Add description of code version for bug report by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/745
* [misc] fix rfc template by @hsliuustc0106 in https://github.com/vllm-project/vllm-omni/pull/748
* fix:#issue 432 by @GG-li in https://github.com/vllm-project/vllm-omni/pull/517
* [Diffusion][Feature] Implement SP support in LongCatImageTransformer by @mxuax in https://github.com/vllm-project/vllm-omni/pull/721
* [Debug] Clean code in Qwen 3 Omni and add warning for talker temperature. by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/688
* [feature] cpu offloading support for diffusion by @LawJarp-A in https://github.com/vllm-project/vllm-omni/pull/497
* [Misc] Group omni arguments into OmniConfig section by @fake0fan in https://github.com/vllm-project/vllm-omni/pull/744
* [Misc] Enable tensor_parallel_size argument with online serving cmd by @JustQJ in https://github.com/vllm-project/vllm-omni/pull/761
* [Bugfix] Raise ValueError when joint_strategy='rear' and causal=True in Ring Attention by @mxuax in https://github.com/vllm-project/vllm-omni/pull/767
* [Feat] add vllm-omni version collection by @sihyeonn in https://github.com/vllm-project/vllm-omni/pull/740
* [Doc] refactor diffusion doc by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/753
* [Bugfix] Fix stable diffusion3 compatibility error by @iwzbi in https://github.com/vllm-project/vllm-omni/pull/772
* [Feature] Support Qwen3 Omni talker mtp batch inference by @ZeldaHuang in https://github.com/vllm-project/vllm-omni/pull/722
* [BugFix]Remove duplicate error handling for request results by @liuyuhanalex in https://github.com/vllm-project/vllm-omni/pull/781
* [CI] Add pytest markers in config files. by @congw729 in https://github.com/vllm-project/vllm-omni/pull/719
* [Doc] Fix mkdocs. by @congw729 in https://github.com/vllm-project/vllm-omni/pull/785
* [Bugfix] Fix generation artifacts of Qwen-Image-Edit-2511 and update pipeline DiT param parsing  by @SamitHuang in https://github.com/vllm-project/vllm-omni/pull/776
* [bugfix] Fix Wan2.2 I2V warmup failure by adding support_image_input attribute by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/791
* [Misc] add wechat group and star history on README by @david6666666 in https://github.com/vllm-project/vllm-omni/pull/801
* [BugFix] Fix incorrect mrope positions under cuda graph by @ZeldaHuang in https://github.com/vllm-project/vllm-omni/pull/803
* [BugFix] Qwen2.5-omni supress end token and won't stop by @yinpeiqi in https://github.com/vllm-project/vllm-omni/pull/773
* [Feature] Flash Attention to Support Attention Mask by @wtomin in https://github.com/vllm-project/vllm-omni/pull/760
* [Model] add flux2 klein by @david6666666 in https://github.com/vllm-project/vllm-omni/pull/809
* [bugfix] use unipc scheduler for Wan 2.2 by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/804
* [Test] Add full test for Qwen3-Omni-30B-A3B-Instruct by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/720
* [Bagel] Support Cache-Dit by @princepride in https://github.com/vllm-project/vllm-omni/pull/736
* [Perf] Optimize the Qwen2.5-Omni Model thinker-to-talker-proj with nn.Linear by @kechengliu97 in https://github.com/vllm-project/vllm-omni/pull/825
* [Core]Add GPU Diffusion Runner by @princepride in https://github.com/vllm-project/vllm-omni/pull/822
* [Feature]: Add CFG param to online serving by @gDINESH13 in https://github.com/vllm-project/vllm-omni/pull/824
* [diffusion] add tp support for qwen-image and refactor some tests by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/830
* [Core] Implement Diffusion Profiler Support by @lishunyang12 in https://github.com/vllm-project/vllm-omni/pull/709
* [Bugfix] Diffusion model fails to load when stage config is present by @fhfuih in https://github.com/vllm-project/vllm-omni/pull/860
* chore: Bump up cache-dit and fix docs links by @DefTruth in https://github.com/vllm-project/vllm-omni/pull/863
* [Misc] Change benchmark default port by @NickLucche in https://github.com/vllm-project/vllm-omni/pull/872
* Dev/rebase 0.14.0 and Support GLM-Image by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/847
* [Doc] Add user guide for diffusion model profiling by @lishunyang12 in https://github.com/vllm-project/vllm-omni/pull/738
* [Doc] Update rebase doc by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/878
* [Test] Add full test for Qwen3-Omni-30B-A3B-Instruct for image and audio single modal by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/827
* [Diffusion] Non-Intrusive Sequence Parallelism (SP) Model Support Abstraction for vLLM-Omni Framework by @mxuax in https://github.com/vllm-project/vllm-omni/pull/779
* [Bugfix] Remove the duplicate api registration in vllm-omni by @fake0fan in https://github.com/vllm-project/vllm-omni/pull/880
* [CI] split tests to avoid timeout by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/883
* [Diffusion][Acceleration] Support TeaCache for Z-Image by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/817
* [Perf] Fuse Q/K/V Linear with QKVParallelLinear in Qwen2.5Omni DiTAttention by @kechengliu97 in https://github.com/vllm-project/vllm-omni/pull/884
* [bugfix] support text + audio mixed output by @GG-li in https://github.com/vllm-project/vllm-omni/pull/843
* [Misc] fix qwen image family redundant computation by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/868
* [ROCm] [CI] Optimize Dockerfilerocm and reduce build time on CI by @tjtanaa in https://github.com/vllm-project/vllm-omni/pull/885
* [Core]Add Diffusion executor by @natureofnature in https://github.com/vllm-project/vllm-omni/pull/865
* [Bugfix] Fix video saving bug under certain fps by @SamitHuang in https://github.com/vllm-project/vllm-omni/pull/893
* [diffusion] use fa3 by default when device supports it by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/783
* [Model] Support Qwen3-TTS model series by @Gaohan123 in https://github.com/vllm-project/vllm-omni/pull/895
* Support Bagel Model by @princepride in https://github.com/vllm-project/vllm-omni/pull/726
* debug Qwen TTS by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/902
* [ROCm] [Doc] Add instructions to install ROCm dependencies to run Qwen3 TTS model by @tjtanaa in https://github.com/vllm-project/vllm-omni/pull/905
* Bump version to 0.14.0rc1 by @ywang96 in https://github.com/vllm-project/vllm-omni/pull/910
* [Doc] Fix the version in the documentation to v0.14.0rc1; fix dockerfile.rocm entrypoint by @tjtanaa in https://github.com/vllm-project/vllm-omni/pull/913
* [Misc] update wechat image by @david6666666 in https://github.com/vllm-project/vllm-omni/pull/914
* [Test] Add precision test cases for Qwen3-Omni-30B-A3B-Instruct in CI by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/828
* [Misc] Fix error log for the diffusion stage timeout by @SamitHuang in https://github.com/vllm-project/vllm-omni/pull/915
* [bugfix] qwen3-tts check_model_inputs by @qibaoyuan in https://github.com/vllm-project/vllm-omni/pull/924
* [Misc] Fix t2i online serving example by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/928
* [examples] add --enable-cpu-offload args by @david6666666 in https://github.com/vllm-project/vllm-omni/pull/930
* [BugFix] Standardize StableAudio audio output by @LudovicoYIN in https://github.com/vllm-project/vllm-omni/pull/842
* [Bagel] Support TeaCache by @princepride in https://github.com/vllm-project/vllm-omni/pull/848
* [Doc] Add Bagel model support to TeaCache documentation by @nussejzz in https://github.com/vllm-project/vllm-omni/pull/943
* [Bugfix] raise error in diffusion engine and fix offload test by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/933
* [Feature]Support async computation and communication across stages by chunks by @amy-why-3459 in https://github.com/vllm-project/vllm-omni/pull/727
* [Bugfix] Fix diffusion pipeline CFG (guidance_scale parsing failure bug) by @SamitHuang in https://github.com/vllm-project/vllm-omni/pull/922
* [doc]Add Text-To-Audio Readme documentation by @zzhuoxin1508 in https://github.com/vllm-project/vllm-omni/pull/958
* [Feature] Diffusion LoRA Adapter Support (PEFT compatible) for vLLM alignment by @AndyZhou952 in https://github.com/vllm-project/vllm-omni/pull/758
* [Diffusion][Feature] Non-Intrusive Sequence Parallelism (SP) Support for Wan2.2 by @mxuax in https://github.com/vllm-project/vllm-omni/pull/966
* [doc] add some additional information for the diffusion model support by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/952
* [Perf] avoid cpu op in QwenImageCrossAttention by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/942
* Support Qwen3 tts online serving by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/968
* [test] fix test_image_generation_lora CI timeout by @AndyZhou952 in https://github.com/vllm-project/vllm-omni/pull/975
* [diffusion] add tp for FLUX.2-klein by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/973
* [test] revert test flash attn file by @AndyZhou952 in https://github.com/vllm-project/vllm-omni/pull/972
* [NPU] Upgrade to v0.14.0 by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/820
* [ROCm] [Perf] Add AITER Flash Attention by @tjtanaa in https://github.com/vllm-project/vllm-omni/pull/941
* [Test] Add chunk e2e test case for CI by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/956
* [Feature] Implement YuanrongConnector based on OmniConnectorBase by @yangsonglin13 in https://github.com/vllm-project/vllm-omni/pull/716
* [NFC] Remove redundant torch.no_grad in models and pipelines by @yuanheng-zhao in https://github.com/vllm-project/vllm-omni/pull/854
* [BugFix] Modify the method of obtaining external_request_id by @amy-why-3459 in https://github.com/vllm-project/vllm-omni/pull/961
* Fix TTS speaker typo and add supported languages by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/990
* [Fix] make images LoRA e2e less flaky by @dongbo910220 in https://github.com/vllm-project/vllm-omni/pull/978
* [Docs] Fix GLM-Image docstring indentation to resolve CI failure by @dongbo910220 in https://github.com/vllm-project/vllm-omni/pull/992
* [Perf] replace torch's conv with vLLM's conv to fix torch 2.9 performance regression by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/982
* [BugFix] Add /health and /v1/models endpoints for diffusion mode, fix streaming compatibility by @majiayu000 in https://github.com/vllm-project/vllm-omni/pull/454
* [Bugfix] Add missing dependencies (onnxruntime, sox) for Qwen3-TTS support by @zzhuoxin1508 in https://github.com/vllm-project/vllm-omni/pull/981
* [Frontend][Model] Support batch request with refined OmniDiffusionReq… by @fhfuih in https://github.com/vllm-project/vllm-omni/pull/797
* [Model]: add FLUX.1-dev model by @nuclearwu in https://github.com/vllm-project/vllm-omni/pull/853
* [BugFix] ignore mm data from stages to async omni by @Bounty-hunter in https://github.com/vllm-project/vllm-omni/pull/954
* Revert "[BugFix] ignore mm data from stages to async omni" by @hsliuustc0106 in https://github.com/vllm-project/vllm-omni/pull/1023
* [Bugfix] Modify output to model_runner_output by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1026
* [Feature] Support cache-dit for Wan 2.2 inference by @SamitHuang in https://github.com/vllm-project/vllm-omni/pull/1021
* [Doc]Format profiling doc by @lishunyang12 in https://github.com/vllm-project/vllm-omni/pull/993
* [Hardware] Support platforms and plugin system by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/774
* [Core]: KV Cache Transfer Encapsulation by @princepride in https://github.com/vllm-project/vllm-omni/pull/979
* [Test]Delete skip mark for amd ci test and fix CI failure by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/927
* [Bugfix][Doc]Specify Qwen3-TTS model name for each task type by @kylehh in https://github.com/vllm-project/vllm-omni/pull/1036
* [Misc] pin version of fa3-fwd by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/1051
* [CI] [ROCm] Add more AMD CI tests by @tjtanaa in https://github.com/vllm-project/vllm-omni/pull/1039
* [Bugfix] fix qwen image layerd in dummy run by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/1027
* [BugFix] Fix noisy output without setting a seed in Qwen Image by @natureofnature in https://github.com/vllm-project/vllm-omni/pull/1043
* [bugfix] remove vllm speech route by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/1060
* [Debug] Update GLM-Image Pipeline by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1049
* [Diffusion][Bugfix] Fix the flash_attn backends selection logic by @mxuax in https://github.com/vllm-project/vllm-omni/pull/983
* [BugFix] Fix the accuracy issue of multimodal input. by @amy-why-3459 in https://github.com/vllm-project/vllm-omni/pull/1020
* [Bugfix] Set VaeImageProcessor `do_convert_rgb` True by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1032
* [feat]: adapt batch request for flux by @nuclearwu in https://github.com/vllm-project/vllm-omni/pull/1028
* [CI] Change Qwen3 Omni stage placement strategy  by @ZeldaHuang in https://github.com/vllm-project/vllm-omni/pull/1072
* [BugFix] Fix to use correct attn backend by @divyanshsinghvi in https://github.com/vllm-project/vllm-omni/pull/1038
* [Perf] Qwen3 Omni talker mtp optimization by @ZeldaHuang in https://github.com/vllm-project/vllm-omni/pull/1005
* [Wan2.2] Optimize memory usage with conditional transformer loading by @faaany in https://github.com/vllm-project/vllm-omni/pull/980
* [Feat] Support XPU Backend in vLLM-Omni by @faaany in https://github.com/vllm-project/vllm-omni/pull/191
* [Fix] stabilize diffusion images LoRA E2E across CI drift by @dongbo910220 in https://github.com/vllm-project/vllm-omni/pull/1075
* [Bugfix][Test] Re-enable the log simple tests by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1065
* [Bugfix] pr conflict fix, bugfix ignore mm data from stages to async omni by @Bounty-hunter in https://github.com/vllm-project/vllm-omni/pull/1025
* [Doc][Bagel] Add BAGEL-7B-MoT documentation and edit the default stage configuration by @nussejzz in https://github.com/vllm-project/vllm-omni/pull/987
* [Fix] Increase max wait time for server readiness to accommodate model loading by @AndyZhou952 in https://github.com/vllm-project/vllm-omni/pull/1089
* [Benchmark] Add vLLM-Omni Omni model online benchmark by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/780
* [Bugfix] Remove Mooncake/Yuanrong connector import warning by @natureofnature in https://github.com/vllm-project/vllm-omni/pull/1091
* fix: UnboundLocalError for role in streaming audio/image responses by @PierreLeGuen in https://github.com/vllm-project/vllm-omni/pull/784
* [Misc] update wechat image by @david6666666 in https://github.com/vllm-project/vllm-omni/pull/1096
* [Feature] Support DiT Layerwise (Blockwise) CPU Offloading by @yuanheng-zhao in https://github.com/vllm-project/vllm-omni/pull/858
* [BugFix] Modify max_tokens and modify the log and fix #1103 by @amy-why-3459 in https://github.com/vllm-project/vllm-omni/pull/1097
* [BugFix] Fix modulate_index shape error in Qwen-Image-Edit Task by @mxuax in https://github.com/vllm-project/vllm-omni/pull/1100
* [Platform] Add supports_torch_inductor interface by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1108
* [BugFix] Fix Qwen3 Omni talker mtp torch.compile startup error by @ZeldaHuang in https://github.com/vllm-project/vllm-omni/pull/1104
* [Bugfix] fix request_id of image generation in api server by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/1112
* [Perf]: CFG parallel abstraction by @wtomin in https://github.com/vllm-project/vllm-omni/pull/851
* [BugFix] Fix Qwen3 TTS 0.6B profile run hang (#995) by @marksverdhei in https://github.com/vllm-project/vllm-omni/pull/1082
* [CI] [ROCm] Quick fix amd ci by @tjtanaa in https://github.com/vllm-project/vllm-omni/pull/1116
* [Bugfix] fix benchmark audio timing error and add benchmark test by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/1109
* [Bugfix][Qwen3TTS] Load speaker_id/voices from model configuration by @JuanPZuluaga in https://github.com/vllm-project/vllm-omni/pull/1079
* [NPU] Align with GPUModelRunner by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1114
* [FEATURE] /v1/images/edit interface by @Bounty-hunter in https://github.com/vllm-project/vllm-omni/pull/1101
* [Bugfix] Fix NPU SDPA attention mask shape and semantics by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1031

## New Contributors
* @ApsarasX made their first contribution in https://github.com/vllm-project/vllm-omni/pull/689
* @dongbo910220 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/708
* @asukaqaq-s made their first contribution in https://github.com/vllm-project/vllm-omni/pull/529
* @nuclearwu made their first contribution in https://github.com/vllm-project/vllm-omni/pull/718
* @yenuo26 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/745
* @GG-li made their first contribution in https://github.com/vllm-project/vllm-omni/pull/517
* @JustQJ made their first contribution in https://github.com/vllm-project/vllm-omni/pull/761
* @sihyeonn made their first contribution in https://github.com/vllm-project/vllm-omni/pull/740
* @liuyuhanalex made their first contribution in https://github.com/vllm-project/vllm-omni/pull/781
* @kechengliu97 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/825
* @gDINESH13 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/824
* @NickLucche made their first contribution in https://github.com/vllm-project/vllm-omni/pull/872
* @LudovicoYIN made their first contribution in https://github.com/vllm-project/vllm-omni/pull/842
* @nussejzz made their first contribution in https://github.com/vllm-project/vllm-omni/pull/943
* @amy-why-3459 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/727
* @zzhuoxin1508 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/958
* @AndyZhou952 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/758
* @yangsonglin13 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/716
* @majiayu000 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/454
* @kylehh made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1036
* @PierreLeGuen made their first contribution in https://github.com/vllm-project/vllm-omni/pull/784
* @marksverdhei made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1082
* @JuanPZuluaga made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1079

**Full Changelog**: https://github.com/vllm-project/vllm-omni/compare/v0.12.0rc1...v0.14.0