[Klaud Cold] Update kimik2.5-int4-h100-vllm vLLM image to v0.22.0

[functionstackx]

Summary

Update vLLM image from v0.20.2 to v0.22.0

Recipes touched: kimik2.5-int4-h100-vllm

Test plan

• full-sweep-enabled sweep passes.

🤖 Generated with Claude Code

---

Note

Medium Risk
Image-only change for a benchmark recipe, but a major vLLM version jump can shift inference behavior and sweep results until full-sweep validation passes.

Overview
Bumps the vLLM container for kimik2.5-int4-h100-vllm from vllm/vllm-openai:v0.20.2 to vllm/vllm-openai:v0.22.0 in .github/configs/nvidia-master.yaml. Scenario coverage is unchanged (still agentic-coding only; no new fixed-seq-len cases).

perf-changelog.yaml gains an entry documenting this image pin for the same config key.

^{Reviewed by Cursor Bugbot for commit caf6b82. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

Thanks for the contribution! For vLLM & SGLang, please ensure that your recipes is similar to the official vLLM recipes and/or the SGLang cookbook

• https://docs.sglang.io/cookbook/intro
• https://recipes.vllm.ai/

If it is not, please create a PR first before we can merge your single node PR into the master branch. Let's ensure that the documentation is first class such that the entire ML community can benefit from your hard work! Thank you

• https://github.com/vllm-project/recipes
• https://github.com/sgl-project/sglang/tree/main/docs_new

PR authors are responsible for ensuring that after merging, all GitHub Action jobs fully pass. A lot of the time, failures are just flakes and simply re-running the failed jobs will fix it. If re-running failed jobs is attempted, PR authors are responsible for ensuring it passes. See GitHub's docs on re-running failed jobs: https://docs.github.com/en/actions/how-tos/manage-workflow-runs/re-run-workflows-and-jobs#re-running-failed-jobs-in-a-workflow

As a rule of thumb, generally, PR authors should request a review & get a PR approval from the respective companies' CODEOWNERS before requesting a review from core maintainers.

If additional help is needed, PR authors can reach out to core maintainers over Slack.

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=26696219784
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=26696219784

[claude]

LGTM, straightforward vLLM image version bump for a single recipe.

Extended reasoning...

Overview

This PR bumps the vLLM image for the kimik2.5-int4-h100-vllm recipe from v0.20.2 to v0.22.0 in .github/configs/nvidia-master.yaml, and adds a matching entry to perf-changelog.yaml. No other fields in the recipe (model, runner, args, sweep config) are altered.

Security risks

None. This is a benchmark configuration update — no auth, crypto, permissions, or runtime logic changes. The image reference points to the official vllm/vllm-openai upstream tag.

Level of scrutiny

Low. This is a mechanical dependency pin change for a single benchmark recipe, following an established pattern used throughout the repo. The full-sweep-enabled label indicates the sweep will validate behavior before merge.

Other factors

No bugs were flagged by the bug hunting system. No outstanding reviewer comments. The changelog entry is properly formatted and consistent with surrounding entries.

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=26696221081
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=26696221081