[tzhouam]

Purpose

First cherry-pick batch for release/v0.32.0 (cut from aa3fd1d9b, the last commit green on CUDA, AMD, Intel, NPU and release pipelines).

Until v0.32.0 is officially released, eligible PRs merged to main (bugfix, CI, docs, release notes, or v0.32.0 milestone) are cherry-picked to the release branch in batches.

Included (clean git cherry-pick -x, main merge order)

• [Bugfix] Keep coordinator broadcasts scheduled during publish #9172 [Bugfix] Keep coordinator broadcasts scheduled during publish
• [Bugfix][Benchmark][Duplex] Fall back to frame-sample judging when the video exceeds the judge context #9101 [Bugfix][Benchmark][Duplex] Fall back to frame-sample judging when the video exceeds the judge context
• [Bugfix] Import Mooncake connectors lazily so an optional dependency cannot break model inspection #9168 [Bugfix] Import Mooncake connectors lazily so an optional dependency cannot break model inspection
• [Doc] Host model recipes in recipes/models/ #9184 [Doc] Host model recipes in recipes/models/
• [Model] Restore Pi0.5 (lerobot/pi05_base) #9080 [Model] Restore Pi0.5 (lerobot/pi05_base): [Model] Remove Pi0.5, HiDream-I1, and LTX-2-19b distilled #9065 removed Pi0.5 before the rc1 cut; without this, v0.32.0 would drop a model supported in v0.30.0.
• [Bugfix] Optionally bind fa3-fwd dense FlashAttention on Hopper #9201 [Bugfix] Optionally bind fa3-fwd dense FlashAttention on Hopper (fixes [Bug][Perf]: qwen-image, performance metrics regressed by more than 10% compared to the baseline in some scenarios #9092, Qwen-Image Hopper latency regression)

Not included

• [New Model] Ming-3-TTS (inclusionAI/Ming-3.0-tts-tiny) Support #9119 Ming-3-TTS: new model, deferred to the next release.
• [Model] Run MiniCPM-o 4.5 speaker embeddings on CUDA #9178 MiniCPM-o speaker embeddings on CUDA: pending release-manager approval.

Test Plan

PR CI on this branch (includes the restored Pi0.5 unit/parity tests).

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	⚠️ Failed 2026-10-21T07:32:08.177014Z We couldn’t complete this review. Comment @codex review to try again.	df2b8b5	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits.
Credits must be used to enable repository wide code reviews.

[vllm-omni-review-bot]

This PR appears to belong to: docs/design/module/diffusion/parallelism.md.

Module owners: @princepride @Bounty-hunter @xuechendi

Routing: @princepride via module of the changed files, CODEOWNERS; @Bounty-hunter via module of the changed files; @xuechendi via module of the changed files

@tzhouam, please review your own changes and leave a short self-review comment describing what you checked. PRs without author self-review may not be assigned a reviewer.

Please take a look when you have a chance. If you would like an automated review, mention @vllm-omni-review-bot in a comment.

[vllm-omni-review-bot]

Omni ReviewBot triage note

Automated triage of commit 4e4e853ec69d produced:

• Priority: high. Prompt maintainer attention is suggested.

These are automated triage suggestions only — the final decision belongs to the maintainers.