chore: bump torch 2.9.1, vllm 0.15 sglang 0.5.8, ray 2.53

[terrykong]

What does this PR do ?

Add a one line overview of what this PR aims to accomplish.

blocked by #1787 (plan is to update mbridge first to unblock nano and then merge this)

Issues

List issues that this PR closes (syntax):

Usage

• You can potentially add a usage example below

# Add a code snippet demonstrating how to use this

Before your PR is "Ready for review"

Pre checks:

• Make sure you read and followed Contributor guidelines
• Did you write any new necessary tests?
• Did you run the unit tests and functional tests locally? Visit our Testing Guide for how to run tests
• Did you add or update any necessary documentation? Visit our Document Development Guide for how to write, build and test the docs.

Additional Information

• ...

Summary by CodeRabbit

• Chores
  • Updated core framework dependencies to latest stable versions, including enhancements to AI inference and distributed computing systems.
  • Removed deprecated kernel package references from project metadata and configurations.
  • Synchronized runtime environment specifications across deployment configurations for improved system stability and compatibility.

[coderabbitai]

📝 Walkthrough

Walkthrough

This PR updates core Python package dependencies in pyproject.toml, including minor version bumps for torch (2.9.0→2.9.1), ray (2.49.2→2.53.0), vllm (0.11.2→0.15.0), and sglang (0.5.7→0.5.8). It removes or comments out sgl-kernel package references across multiple dependency lists.

Changes

Cohort / File(s)	Summary
Dependency Updates pyproject.toml	Bumped torch, ray, vllm, and sglang versions; removed/commented out sgl-kernel references in multiple dependency sections and metadata; updated override-dependencies for torch and torchaudio versions.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Possibly related PRs

• chore: Bump vllm to 0.11.2, torch to 2.9, transformers to 4.57.1 #1563: Updates the same core dependencies (torch, vllm) in pyproject.toml
• fix: update the custom vllm instructions #1116: Modifies pyproject.toml dependency handling for vllm and related build/workflow configurations

Suggested labels

CI, CI:L1

Suggested reviewers

• yfw
• joyang-nv

🚥 Pre-merge checks | ✅ 3 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Test Results For Major Changes	⚠️ Warning	PR contains substantial dependency updates (vllm 0.11.2→0.15.0, ray 2.49.2→2.53.0) with incomplete description lacking test results and validation.	Include test results validating compatibility with new vllm/ray versions, performance benchmarks confirming no regression, and verify pyproject.toml dependency conflicts are handled.

✅ Passed checks (3 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title directly matches the main changes in the pull request—it lists the specific package versions being bumped (torch 2.9.1, vllm 0.15, sglang 0.5.8, ray 2.53), which are the primary modifications evident in pyproject.toml.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch tk/vllm-torch-bump

Important

Action Needed: IP Allowlist Update

If your organization protects your Git platform with IP whitelisting, please add the new CodeRabbit IP address to your allowlist:

• ✨ 136.113.208.247/32 (new)
• 34.170.211.100/32
• 35.222.179.152/32

Reviews will stop working after February 8, 2026 if the new IP is not added to your allowlist.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🤖 Fix all issues with AI agents

In `@pyproject.toml`:
- Around line 22-25: The build-group entry pins "torch==2.9.0" while the runtime
dependencies list "torch==2.9.1", causing an ABI mismatch; update the
build-group torch pin from "torch==2.9.0" to "torch==2.9.1" so both build and
runtime use the exact same torch version (look for the "torch==2.9.0" string in
the build-group section of pyproject.toml and replace it with "torch==2.9.1").

🧹 Nitpick comments (1)

pyproject.toml (1)

294-335: Clean up or gate remaining sgl-kernel config after disabling it.

sgl-kernel is now commented out in extra-build-dependencies and dependency-metadata, but extra-build-variables still defines sgl-kernel settings (Line 215). Consider removing or gating those entries to avoid stale config and confirm sglang no longer requires sgl-kernel at build/runtime.