v0.16.0

## Highlights
This release features approximately 121 commits (merged PRs) from ~60 contributors (24 new contributors).

vLLM-Omni v0.16.0 is a major alignment + capability release that rebases the project onto **upstream vLLM v0.16.0** and significantly expands performance, distributed execution, and production readiness across **Qwen3-Omni / Qwen3-TTS**, **Bagel**, **MiMo-Audio**, **GLM-Image** and the **Diffusion (DiT) image/video stack**—while also improving platform coverage (CUDA / ROCm / NPU / XPU), CI quality, and documentation. 

### Key Improvements

* **Rebase to upstream vLLM v0.16.0**: Tracks the latest vLLM runtime behavior and APIs while keeping Omni’s error handling aligned with upstream expectations. (#1357, #1122, plus follow-up fixes like #1401)
* **Qwen3-Omni performance + correctness**: Performance optimizations (cuda graph, async-chunk, streaming output) make TTFP reduced 90% and RTF 0.22~0.45, plus precision and E2E metric correctness fixes. (#1378, #1352, #1288, #1018, #1292)
* **MiMo-Audio production Support**: Performance optimizations  (cuda graph, async-chunk, streaming output) improves the RTF ~ 0.2, 11x faster than baseline. #750 
* **Qwen3-TTS production upgrades**: Disaggregated inference pipeline support, streaming output, batched Code2Wav decoding, and CUDA Graph support for speech tokenizer decoding—plus multiple robustness fixes across task type handling and voice cloning. (#1161, #1438, #1426, #1205, #1317, #1554)
* **Bagel acceleration & scalability**: Adds **TP support**, introduces **CFG** capabilities, and accelerates multi-branch CFG by **merging branches into a single batch**; includes KV transfer stability fixes. (#1293, #1310, #1429, #1437)
* **Diffusion distributed execution expansion**: Adds/extends TP/SP/HSDP and reduces redundant communication overhead; improves pipeline parallelism options (e.g., VAE patch parallel) and correctness across multiple diffusion families. (#964, #1275, #1339, #756, #1428)
* **Quantization for DiT**: Introduces **FP8** quantization support and **native GGUF quantization** support for diffusion transformers, with code-path cleanups. (#1034, #1285, #1533)
* **Broader model coverage (audio + image)**: Adds **MiMo-Audio-7B-Instruct** support and performance improvements for **GLM-Image** pipelines. (#750, #920)

---

### Diffusion, Image & Video Generation

* **New/expanded model coverage**

  * **HunyuanImage3** support and v0.16.0 follow-ups removing CUDA hardcoding + MOE fixes. (#1085, #1402, #1401)
  * **OmniGen2** support. (#513)
  * **nextstep_1** diffusion model (T2I-only). (#612)
* **Distributed & parallel execution**

  * **TP support** additions/expansions for diffusion models (e.g., Wan 2.2, SD 3.5). (#964, #1336)
  * **HSDP for diffusion models** for improved scalability. (#1339)
  * **VAE patch parallelism** support (and enablement for SD3.5). (#756, #1428)
  * **Sequence-parallel comm reduction** by refining SP hook design. (#1275)
* **Performance & memory efficiency**

  * Flux caching features (e.g., cache_dit) and **CFG-parallel** improvements for Flux.1-dev. (#1145, #1269)
  * Process-level memory calculation hooks for diffusion workloads. (#1276)
  * Platform-wide enablement of **layerwise offload**. (#1492)
* **Correctness & stability**

  * Multiple pipeline stability and correctness fixes (seed handling, attention mask dtype/shape, tokenizer padding issues, init/download safety, model detect robustness, etc.). (e.g., #1249, #1248, #1349, #1241, #1213, #1254, #1562)

---

### Audio, Speech & Omni (Qwen3-TTS / MiMo-Audio)

* **Qwen3-TTS feature set maturation**

  * Disaggregated inference pipeline support for stage-based / split deployment. (#1161)
  * **Streaming output** for `v1/audio/speech`-style workflows. (#1438)
  * **Code2Wav batched decoding** and async-chunk batch inference enhancements. (#1426, #1246)
  * CUDA Graph support for the speech tokenizer decoder. (#1205)
* **Stability & quality**

  * Fixes for task_type handling, snapshot/download behavior, configuration options, and voice clone corruption edge cases. (#1317, #1318, #1177, #1554, #1455)
  * More robust handling of **multimodal outputs that attach audio payloads** and related server-side audio data processing. (#1203, #1222)

---

### Multimodal Model Improvements

* **Bagel**

  * **TP support** for scaling across devices. (#1293)
  * **CFG enablement** and **multi-branch CFG merged into a single batch** to improve throughput and reduce per-branch overhead. (#1310, #1429)
  * KV transfer and stability fixes. (#1437)
* **GLM-Image**

  * **Performance improvements** for GLM-Image workloads. (#920)
  * Additional image-serving hardening that benefits GLM-Image deployments (endpoint/pipeline validation and crash fixes in edge cases). (e.g., #1141, #1265, #1248)

---

### Serving, APIs & Integrations

* **OpenAI-compatible video serving**

  * Adds **Wan2.2 T2V and I2V online serving** via OpenAI **`/v1/videos`** API. (#1073)
  * Supports irregular output shapes for Wan2.2. (#1279)
* **Online serving robustness & usability**

  * Unify CLI argument naming style and forward serve parameters more consistently to models. (#1309, #985)
  * Per-request `generator_device` for online image generation/edit flows. (#1183)
  * Fixes for image edit endpoint validation and RoPE crashes on explicit H/W. (#1141, #1265)
* **Ecosystem integration**

  * **ComfyUI integration** for improved workflow adoption. (#1113)

---

### Performance, Scheduling & Memory Accounting

* **Async chunk enhancements**

  * Overlap chunk I/O and compute via async scheduling to reduce idle time in chunked pipelines. (#951)
  * Async-chunk refactors and shape mismatch fixes for stability. (#1151, #1195)
* **Metrics & benchmarking**

  * Metrics structure optimization and multiple fixes for token/stream stats and E2E correctness (including Qwen3-Omni async-chunk E2E metric correctness). (#891, #1292, #1301, #1018)
  * Adds benchmarks for audio speech non-streaming and omni performance benchmark tests. (#1408, #1321)
* **Memory accounting**

  * **Process-scoped GPU memory accounting** and diffusion-side process-level tracking improvements. (#1204, #1276)

---

### Platform, Hardware Backends & Deployment

* **XPU / NPU / ROCm coverage improvements**

  * XPU Dockerfile + docs, enable FLASH_ATTN on XPU, fix XPU UT coverage; disable diffusion compile on XPU where needed. (#1162, #1332, #1164, #1148)
  * NPU upgrade to v0.16.0 and recovery fixes for Qwen3-TTS. (#1375, #1564)
  * ROCm CI/docker updates to track vLLM v0.16.0 stable. (#1380, #1500)
* **Deployment & connectivity**

  * Stage-based deployment CLI and UDS-based ZMQ address handling for stage serving. (#939, #1522)
  * **RDMA connector** support for high-performance interconnect scenarios. (#1019)
  * Platform-dependent package installation improvements. (#1046)

---

### CI, Testing, Docs & Developer Experience

* **CI quality + coverage**

  * Expanded test stratification design (L2/L3), nightly(L4) test runs, branch coverage fixes, and CI performance tuning. (#1272, #1333, #1120, #1283)
  * Improved CI stability (timeouts, reduced H100 usage, clearer logs). (#1460, #1543, #1463)
* **Docs & tutorials**

  * Tutorials on models/pipelines/features, diffusion tutorial refinements, Qwen3-TTS docs consistency, quantization Q&A updates, and installation instructions for vLLM 0.16.0. (#1196, #1305, #1226, #1257, #1505)
  * Improved examples (e.g., image-to-video download steps). (#1258)
* **Tooling**

  * Online profiling support and other developer ergonomics improvements. (#1136)

---

### Stability & Bug Fixes (Across the Stack)

This release includes broad correctness and robustness fixes spanning:

* **Diffusion pipelines** (dtype/shape, init crashes, model detection, seed and config handling)
* **Image edit / generation** endpoints (format validation, RoPE crash, argument typing, seed handling)
* **Distributed execution** (process group mapping accuracy, scheduler race conditions, kv transfer correctness)
* **General runtime hygiene** (removing unnecessary ZMQ init, CLI naming normalization, upstream-aligned error handling)

## What's Changed
* [TeaCache]: Add Coefficient Estimation by @princepride in https://github.com/vllm-project/vllm-omni/pull/940
* [CI]: Bagel E2E Smoked Test by @princepride in https://github.com/vllm-project/vllm-omni/pull/1074
* [Misc] Bump version to 0.14.0 by @ywang96 in https://github.com/vllm-project/vllm-omni/pull/1128
* [Doc] First stable release of vLLM-Omni by @ywang96 in https://github.com/vllm-project/vllm-omni/pull/1129
* [Misc] Align error handling with upstream vLLM v0.14.0 by @ceanna93 in https://github.com/vllm-project/vllm-omni/pull/1122
* [Feature] add Tensor Parallelism to LongCat-Image(-Edit) by @hadipash in https://github.com/vllm-project/vllm-omni/pull/926
* [CI] Temporarily remove slow tests. by @congw729 in https://github.com/vllm-project/vllm-omni/pull/1143
* [CI] Refactor test_sequence_parallel.py and add a warmup run for more accurate performance stat by @mxuax in https://github.com/vllm-project/vllm-omni/pull/1165
* Dev/rebase v0.15.0 by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1159
* Docs update paper link by @hsliuustc0106 in https://github.com/vllm-project/vllm-omni/pull/1169
* [Debug] Clear Dockerfile.ci to accelerate build image by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1172
* [Debug] Correct Unreasonable Long Timeout by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1175
* [Doc]Fix - Align with repo. by @congw729 in https://github.com/vllm-project/vllm-omni/pull/1176
* [Bugfix][Qwen-Image-Edit] Add a warning log for none negative_prompt by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1170
* [Bugfix] fix qwen image oom by @ZJY0516 in https://github.com/vllm-project/vllm-omni/pull/1168
* [Hardware] Disable compile of diffusion on XPU by @zhenwei-intel in https://github.com/vllm-project/vllm-omni/pull/1148
* [Doc] Fix vLLM version in user docs by @yuanheng-zhao in https://github.com/vllm-project/vllm-omni/pull/1179
* [Refactor] Refactor async chunk and fix the shape mismatch issue by @amy-why-3459 in https://github.com/vllm-project/vllm-omni/pull/1151
* bugfix: /images/edits endpoint fails pipeline data format check by @fhfuih in https://github.com/vllm-project/vllm-omni/pull/1141
* [Perf] resolving prolonged `cudastreamsynchronize` execution in z image processing by @erfgss in https://github.com/vllm-project/vllm-omni/pull/1105
* [Bugfix] modify RTF use audio_e2e/audio_duration by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/1157
* [Doc] Highlight paper & slides. by @congw729 in https://github.com/vllm-project/vllm-omni/pull/1186
* [chore] Remove zmq context initialize by @xiedeyantu in https://github.com/vllm-project/vllm-omni/pull/1187
* [NPU] Update Dockerfile and docs for v0.14.0 by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/671
* [Bugfix] E2E metric incorrect qwen3-omni with async chunk feature by @LJH-LBJ in https://github.com/vllm-project/vllm-omni/pull/1018
* [Doc] opt doc by @david6666666 in https://github.com/vllm-project/vllm-omni/pull/1118
* [Bugfix] Fix tp+sp accuracy, incorrect process group mapping by @david6666666 in https://github.com/vllm-project/vllm-omni/pull/1178
* [Feature] Enable use_audio_in_video for Qwen 3 Omni Online by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1198
* [Bugfix] async_chunk rebase v0.15.0 by @amy-why-3459 in https://github.com/vllm-project/vllm-omni/pull/1195
* [feature]: support flux cache_dit by @nuclearwu in https://github.com/vllm-project/vllm-omni/pull/1145
* [CI] Add CI branch coverage calculation,  fix statement coverage results and add log before test for buildkite  log group by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/1120
* [Wan 2.2][Diffusion] Add TP Support by @Pr0Wh1teGivee in https://github.com/vllm-project/vllm-omni/pull/964
* [Hardware] [Feat] Setup platform dependent package installation by @tjtanaa in https://github.com/vllm-project/vllm-omni/pull/1046
* [XPU] Fix XPU UTs for basic coverage by @yma11 in https://github.com/vllm-project/vllm-omni/pull/1164
* [Test] Add BuildKite test-full script for full CI. by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/867
* [Refactor] Reuse upstream Qwen3MoeSparseMoeBlock by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1202
* [Bugfix] Fix wan2.2 ti2v by @mxuax in https://github.com/vllm-project/vllm-omni/pull/1221
* [Bugfix] Fix '--max-generated-image-size' cli args type by @ApsarasX in https://github.com/vllm-project/vllm-omni/pull/1249
* [Bugfix] Ensure seed=0 is correctly handled in image edit by @ApsarasX in https://github.com/vllm-project/vllm-omni/pull/1248
* [Docs] Add example image download step to Image-To-Video examples by @lishunyang12 in https://github.com/vllm-project/vllm-omni/pull/1258
* [Bugfix] Fix padding bug in 12Hz tokenizer ConvTranspose1d decode by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/1241
* [bugfix] Fix multimodal_output property to check completion outputs where audio data is attached by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/1203
* [Doc] Update QA relevant to quantization  by @lishunyang12 in https://github.com/vllm-project/vllm-omni/pull/1257
* [Bugfix] Fix Doc link Rrror by @lishunyang12 in https://github.com/vllm-project/vllm-omni/pull/1263
* Process-Scoped GPU Memory Accounting by @divyanshsinghvi in https://github.com/vllm-project/vllm-omni/pull/1204
* [ComfyUI]: ComfyUI integration by @fhfuih in https://github.com/vllm-project/vllm-omni/pull/1113
* fix: add diffusion offload args to OmniConfig group instead of serve_parser by @fake0fan in https://github.com/vllm-project/vllm-omni/pull/1271
* [Doc] Adding models/pipelines/features Tutorial by @wtomin in https://github.com/vllm-project/vllm-omni/pull/1196
* [CI] Add env variable check for nightly CI  by @congw729 in https://github.com/vllm-project/vllm-omni/pull/1281
* [CI] Add pytest markers to current tests and update the doc. by @congw729 in https://github.com/vllm-project/vllm-omni/pull/577
* [Diffusion][Perf] Remove Redundant Communication Cost by Refining SP Hook Design by @mxuax in https://github.com/vllm-project/vllm-omni/pull/1275
* [Feature] Opt metrics structure by @LJH-LBJ in https://github.com/vllm-project/vllm-omni/pull/891
* [Test] Add example test cases for omni online by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/1086
* [CI] Reduce the time for Diffusion Sequence Parallelism Test by @congw729 in https://github.com/vllm-project/vllm-omni/pull/1283
* [Model] SupportHunyuanImage3 Diffusion Model in vllm-omni by @ElleElleWu in https://github.com/vllm-project/vllm-omni/pull/1085
* [Chore] Update copyright year. by @lishunyang12 in https://github.com/vllm-project/vllm-omni/pull/1256
* [feature]: support Flux.1-dev CFG-Parallel by @nuclearwu in https://github.com/vllm-project/vllm-omni/pull/1269
* [Bugfix] Fix 'NoneType' AttributeError in stable-diffusion model detect by @yma11 in https://github.com/vllm-project/vllm-omni/pull/1254
* [Doc] Update Qwen3-TTS docs for consistency with Omni examples by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/1226
* [Fix]Ensure HuggingFace downloads complete before initialization. by @zzhuoxin1508 in https://github.com/vllm-project/vllm-omni/pull/1213
* [BugFix] Fixed the issue where ignore_eos was not working. by @amy-why-3459 in https://github.com/vllm-project/vllm-omni/pull/1286
* [Test] Add e2e tests for Qwen3-TTS speech endpoint by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/1206
* [Feat]: support VAE patch parallelism by @dongbo910220 in https://github.com/vllm-project/vllm-omni/pull/756
* [CI] Disable Qwen3-TTS E2E Test in pipeline.yml by @Gaohan123 in https://github.com/vllm-project/vllm-omni/pull/1306
* [Misc] Add per-request generator_device to online image gen and edit by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1183
* [Bagel]: Support TP by @princepride in https://github.com/vllm-project/vllm-omni/pull/1293
* [Bugfix] Fix image edit RoPE crash when explicit height/width are provided by @lishunyang12 in https://github.com/vllm-project/vllm-omni/pull/1265
* [Doc] Sync by @congw729 in https://github.com/vllm-project/vllm-omni/pull/1216
* [Bugfix] fix precision issues of qwen3-omni when enable async_chunk without system prompt by @R2-Y in https://github.com/vllm-project/vllm-omni/pull/1288
* [Debug] Add trigger to concurrent stage init by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1274
* [Bugfix][Qwen3-TTS] Fix task type by @ekagra-ranjan in https://github.com/vllm-project/vllm-omni/pull/1317
* Unifying CLI Argument Naming Style by @wtomin in https://github.com/vllm-project/vllm-omni/pull/1309
* [Bugfix][Qwen3-TTS] Preserve original model ID in omni_snapshot_download by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/1318
* [CI] Run nightly tests. by @congw729 in https://github.com/vllm-project/vllm-omni/pull/1333
* [Feature]: FP8 Quantization Support for DiT  by @lishunyang12 in https://github.com/vllm-project/vllm-omni/pull/1034
* Fix yield token metrics and opt metrics record stats by @LJH-LBJ in https://github.com/vllm-project/vllm-omni/pull/1292
* [Test] L2 & L3 Test Case Stratification Design for Omni Model by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/1272
* [Pref] Support Qwen3 Omni code2wav batch infernce with async chunk by @ZeldaHuang in https://github.com/vllm-project/vllm-omni/pull/1246
* update qwen3-omni & qwen2.5-omni openai client by @R2-Y in https://github.com/vllm-project/vllm-omni/pull/1304
* [Feature] Support Wan2.2 T2V and I2V Online Serving with OpenAI /v1/videos API by @SamitHuang in https://github.com/vllm-project/vllm-omni/pull/1073
* [Feature] add Tensor Parallelism to SD_3.5 by @GG-li in https://github.com/vllm-project/vllm-omni/pull/1336
* [Feature]async scheduling to overlap chunk IO and compute by @Shirley125 in https://github.com/vllm-project/vllm-omni/pull/951
* [Bugfix] reused metrics to modify the API Server token statistics in Stream Response by @kechengliu97 in https://github.com/vllm-project/vllm-omni/pull/1301
* Refactor CPU Offloading Backend Pattern by @yuanheng-zhao in https://github.com/vllm-project/vllm-omni/pull/1223
* [DOC] Doc for CI test - Details about five level stucture and some other files. by @congw729 in https://github.com/vllm-project/vllm-omni/pull/1167
* [Bugfix] remove Tongyi-MAI/Z-Image-Turbo related test from L2 ci by @Bounty-hunter in https://github.com/vllm-project/vllm-omni/pull/1348
* [Misc] wechat image update by @david6666666 in https://github.com/vllm-project/vllm-omni/pull/1354
* [Misc] Support WorkerWrapperBase and CustomPipeline for Diffusion Worker by @knlnguyen1802 in https://github.com/vllm-project/vllm-omni/pull/764
* [Feature][Bugfix] Add CFG feature to Bagel by @nussejzz in https://github.com/vllm-project/vllm-omni/pull/1310
* [Feature]: Diffusion sleep to use process level memory calculation by @divyanshsinghvi in https://github.com/vllm-project/vllm-omni/pull/1276
* change qwen3-omni open cudagraph by default by @R2-Y in https://github.com/vllm-project/vllm-omni/pull/1352
* [XPU] Update Bagel's flash_attn_varlen_func to fa utils by @zhenwei-intel in https://github.com/vllm-project/vllm-omni/pull/1295
* [Test] Add Omni Model Performance Benchmark Test by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/1321
* [BugFix]: Revert utils change by @princepride in https://github.com/vllm-project/vllm-omni/pull/1369
* [Rebase] Rebase to vllm v0.16.0 by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1357
* [Test] Fix expansion and example test case for qwen3-omni by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/1358
* [v0.16.0][BUG FIX]Fix hunyuan MOE after update to 0.16.0 by @xuechendi in https://github.com/vllm-project/vllm-omni/pull/1401
* [0.16.0] remove cuda hard-code for Hunyuan Image3 by @xuechendi in https://github.com/vllm-project/vllm-omni/pull/1402
* [XPU] Add XPU Dockerfile and related docs by @yma11 in https://github.com/vllm-project/vllm-omni/pull/1162
* [Bugfix] Fix Hardcoded Datatypes in Z-image by @alex-jw-brooks in https://github.com/vllm-project/vllm-omni/pull/1393
* [Feature] : Support disaggregated inference pipeline for Qwen3_TTS by @Sy0307 in https://github.com/vllm-project/vllm-omni/pull/1161
* [Feature] Add automated PR reviewer bot with GLM integration by @hsliuustc0106 in https://github.com/vllm-project/vllm-omni/pull/1424
* [Misc] Add Qwen2.5-Omni-3B model support to Gradio demo by @UsamaKenway in https://github.com/vllm-project/vllm-omni/pull/1382
* [misc] Feature/pr reviewer auto trigger&update model by @hsliuustc0106 in https://github.com/vllm-project/vllm-omni/pull/1431
* Revert "[misc] Feature/pr reviewer auto trigger&update model" by @hsliuustc0106 in https://github.com/vllm-project/vllm-omni/pull/1432
* [Doc] Update GPU installation commands by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1434
* [ROCM] [CI] fix dockerfile.rocm to support nightly build and also fix amd ci v0.16.0rc1 by @tjtanaa in https://github.com/vllm-project/vllm-omni/pull/1380
* [Feature][BAGEL] Combine multi-branch cfg into a single batch to accelerate inference. by @nussejzz in https://github.com/vllm-project/vllm-omni/pull/1429
* [Feat]: add ASCII art logo for vLLM-Omni  by @zzhuoxin1508 in https://github.com/vllm-project/vllm-omni/pull/1430
* [Bug] [Bagel] Fix kv transfer bug by @nussejzz in https://github.com/vllm-project/vllm-omni/pull/1437
* [CI] Set L2 & L3 tests running conditions. by @congw729 in https://github.com/vllm-project/vllm-omni/pull/1344
* [Feature] vLLM-Omni RDMA connector by @natureofnature in https://github.com/vllm-project/vllm-omni/pull/1019
* [Minor][Refactor] Pass seq_token_counts explicitly by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1425
* [Misc] Extend Diffusion Benchmark script to other backends by @NickLucche in https://github.com/vllm-project/vllm-omni/pull/875
* [Feature] Support Stage Based Deployment CLI by @wuhang2014 in https://github.com/vllm-project/vllm-omni/pull/939
* [Doc] Optimize vLLM-Omni metrics documentation by @LJH-LBJ in https://github.com/vllm-project/vllm-omni/pull/1311
* [Bugfix]  Forward all vllm-omni serve command parameters to model by @LJH-LBJ in https://github.com/vllm-project/vllm-omni/pull/985
* [Doc]: Add bagel single/multi node usage with mooncake document by @princepride in https://github.com/vllm-project/vllm-omni/pull/1450
* [Qwen3TTS][Feat] Code2Wav batched decoding by @JuanPZuluaga in https://github.com/vllm-project/vllm-omni/pull/1426
* [CI] Remove overwhelming debug log by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1463
* [Misc] update wechat image by @david6666666 in https://github.com/vllm-project/vllm-omni/pull/1464
* [Doc] Refine Diffusion Tutorial Documents by @wtomin in https://github.com/vllm-project/vllm-omni/pull/1305
* [Bugfix] Robust Audio Data Handling in _create_audio_choice by @LJH-LBJ in https://github.com/vllm-project/vllm-omni/pull/1222
* [Bugfix]: Fix merging updated additional information to ensure dict type by @Dovis01 in https://github.com/vllm-project/vllm-omni/pull/1296
* [Model]Add new nextstep_1(Diffusion) model(only T2I) by @sniper35 in https://github.com/vllm-project/vllm-omni/pull/612
* [Bugfix] Add TTS configuration options by @YanickSchraner in https://github.com/vllm-project/vllm-omni/pull/1177
* [Debug] Multi-Request for Qwen 3 Omni use_audio_in_video by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1433
* [Bugfix] Fix case-sensitive task_type matching in Qwen3TTSModelForGeneration by @upskyy in https://github.com/vllm-project/vllm-omni/pull/1455
* [BugFix] process request.num_cached_tokens if it equals to the initial value  by @LJH-LBJ in https://github.com/vllm-project/vllm-omni/pull/1468
* [Bugfix] Fix SDPA attention mask dtype and shape (Fix #857) by @yJader in https://github.com/vllm-project/vllm-omni/pull/1349
* [Test] Reduce Perf test case and fix modify stage config by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/1449
* [NPU] Upgrade to v0.16.0 by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1375
* [CI] Update Dockerfile for vllm-omni CI image and remove obsolete dep… by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1491
* [Fix][Chore] Qwen3-TTS Modeling Minor Code Sanity Improvements by @yuanheng-zhao in https://github.com/vllm-project/vllm-omni/pull/1482
* [Bugfix] Fix tuple/list KV cache extraction crash by @junuxyz in https://github.com/vllm-project/vllm-omni/pull/1405
* [Doc] format lora related docs for the user's end by @AndyZhou952 in https://github.com/vllm-project/vllm-omni/pull/1009
* [Feature] Support Wan2.2 output with irregular shapes by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1279
* [Misc] Migrate L1 tests to use pytest-mock by @yuanheng-zhao in https://github.com/vllm-project/vllm-omni/pull/1315
* [Bugfix] Fix LoRA Scaling on Active Adapters by @alex-jw-brooks in https://github.com/vllm-project/vllm-omni/pull/1421
* [Bugfix] fix record audio generated frame in offline infer by @LJH-LBJ in https://github.com/vllm-project/vllm-omni/pull/1312
* [Model] Support OmniGen2 by @legitnull in https://github.com/vllm-project/vllm-omni/pull/513
* [Bugfix][Qwen3TTS] by @JuanPZuluaga in https://github.com/vllm-project/vllm-omni/pull/1289
* Use pull through cache image for H100 pool by @khluu in https://github.com/vllm-project/vllm-omni/pull/1518
* [ROCm] [CI] [Docker] Point to use the latest vLLM v0.16.0 stable version by @tjtanaa in https://github.com/vllm-project/vllm-omni/pull/1500
* [Bugfix] fix offline text_to_image error from #1009 by @david6666666 in https://github.com/vllm-project/vllm-omni/pull/1515
* [XPU] Enable FLASH_ATTN on XPU by @yma11 in https://github.com/vllm-project/vllm-omni/pull/1332
* Revert gpu_1 job to use regular image by @khluu in https://github.com/vllm-project/vllm-omni/pull/1521
* [Chore] remove unused logger in omni_diffusion (#531) by @fhfuih in https://github.com/vllm-project/vllm-omni/pull/1509
* [Qwen3TTS][Feat] Streaming output by @JuanPZuluaga in https://github.com/vllm-project/vllm-omni/pull/1438
* [Bugfix] Race condition in MultiprocExecutor when concurent access to Scheduler by @knlnguyen1802 in https://github.com/vllm-project/vllm-omni/pull/1448
* [Doc][Test][Misc] ComfyUI test, more screenshot, and code cleaning by @fhfuih in https://github.com/vllm-project/vllm-omni/pull/1435
* [Performance]Qwen3-Omni performance optimization by @amy-why-3459 in https://github.com/vllm-project/vllm-omni/pull/1378
* [Feature] Support HSDP for diffusion models by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1339
* [CI] fixed CI timeout by @zhumingjue138 in https://github.com/vllm-project/vllm-omni/pull/1460
* [Bugfix] Use uds for zmq address if not set --stage-id by @wuhang2014 in https://github.com/vllm-project/vllm-omni/pull/1522
* [BugFix] Restore talker's config by @amy-why-3459 in https://github.com/vllm-project/vllm-omni/pull/1524
* [XPU] fix qwen_omni after rebase to v0.16.0 by @xuechendi in https://github.com/vllm-project/vllm-omni/pull/1416
* [Platform] Enable layerwise offload on all hardware by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1492
* diffusion: enable VAE patch parallel for SD3.5 by @dongbo910220 in https://github.com/vllm-project/vllm-omni/pull/1428
* [Perf] GLM Image by @JaredforReal in https://github.com/vllm-project/vllm-omni/pull/920
* [skip ci][Doc] add design docs for async chunk in qwen3-omni by @R2-Y in https://github.com/vllm-project/vllm-omni/pull/962
* feat(qwen3-tts): Add CUDA Graph support for speech tokenizer decoder by @xulusjb in https://github.com/vllm-project/vllm-omni/pull/1205
* [New Model]: XiaomiMiMo/MiMo-Audio-7B-Instruct support by @qibaoyuan in https://github.com/vllm-project/vllm-omni/pull/750
* [Feature]: Native GGUF Quantization Support for DiT by @david6666666 in https://github.com/vllm-project/vllm-omni/pull/1285
* Add benchmark for `v1/audio/speech` non-streaming by @ekagra-ranjan in https://github.com/vllm-project/vllm-omni/pull/1408
* [Version] Auto generate version using `setuptool_scm` by @tjtanaa in https://github.com/vllm-project/vllm-omni/pull/1224
* [Feat] : Support Async chunk cleanup by @Sy0307 in https://github.com/vllm-project/vllm-omni/pull/1087
* [Profiler] Support online profiling by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1136
* [Bugfix] Fix redundant finished req status updating on OmniGenerationScheduler by @Dovis01 in https://github.com/vllm-project/vllm-omni/pull/1510
* [XPU][NPU][ROCM] enable cpu_offloading flag for non_cuda by @xuechendi in https://github.com/vllm-project/vllm-omni/pull/1488
* [Chore] Cleanup dead code in GGUF DiT code path by @Isotr0py in https://github.com/vllm-project/vllm-omni/pull/1533
* [Doc] Update installation instructions for vllm 0.16.0 by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1505
* [Doc] [skip ci]Sync. by @congw729 in https://github.com/vllm-project/vllm-omni/pull/1363
* [CI][skip ci]Update H100 image link based on #1518 by @congw729 in https://github.com/vllm-project/vllm-omni/pull/1538
* Fix no embed text spk tokens by @LJH-LBJ in https://github.com/vllm-project/vllm-omni/pull/1540
* [Debug] Merge vllm pull 35368 by @tzhouam in https://github.com/vllm-project/vllm-omni/pull/1534
* [Docs] update async chunk docs diagram [skip ci] by @R2-Y in https://github.com/vllm-project/vllm-omni/pull/1530
* fix(qwen3-tts): fix Base ICL voice clone producing corrupted audio by @linyueqian in https://github.com/vllm-project/vllm-omni/pull/1554
* [NPU][Bugfix] Align GPU side and recover qwen3-tts by @gcanlin in https://github.com/vllm-project/vllm-omni/pull/1564
* [BugFix] Fix unexpected crash when init OmniDiffusion by @Semmer2 in https://github.com/vllm-project/vllm-omni/pull/1562
* [CI] Modify some CI test cases to run on L4 environment to reduce H100 resource usage. by @yenuo26 in https://github.com/vllm-project/vllm-omni/pull/1543
* [BugFix]: fix a lot of bug by @princepride in https://github.com/vllm-project/vllm-omni/pull/1565

## New Contributors
* @ceanna93 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1122
* @hadipash made their first contribution in https://github.com/vllm-project/vllm-omni/pull/926
* @zhenwei-intel made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1148
* @erfgss made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1105
* @xiedeyantu made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1187
* @Pr0Wh1teGivee made their first contribution in https://github.com/vllm-project/vllm-omni/pull/964
* @yma11 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1164
* @ElleElleWu made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1085
* @ekagra-ranjan made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1317
* @Shirley125 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/951
* @xuechendi made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1401
* @alex-jw-brooks made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1393
* @Sy0307 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1161
* @UsamaKenway made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1382
* @Dovis01 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1296
* @YanickSchraner made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1177
* @upskyy made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1455
* @yJader made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1349
* @junuxyz made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1405
* @legitnull made their first contribution in https://github.com/vllm-project/vllm-omni/pull/513
* @khluu made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1518
* @zhumingjue138 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1460
* @xulusjb made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1205
* @Semmer2 made their first contribution in https://github.com/vllm-project/vllm-omni/pull/1562

**Full Changelog**: https://github.com/vllm-project/vllm-omni/compare/v0.14.0...v0.16.0