Add B300 config: kimi-k2.5-int4-vllm (vLLM 0.20.0 + TP=4/EP=1 sweep)

[functionstackx]

Note

Supersedes #1071 (which itself reopened #1057 after the #1070 revert). Opening fresh from main since the original branch's merge base had drifted — the b200 schema migrated from seq-len-configs to scenarios.fixed-seq-len between then and now — and a clean reopen reads more cleanly than a rebase.

Summary

• Add kimik2.5-int4-b300-vllm benchmark config and the corresponding benchmarks/single_node/kimik2.5_int4_b300.sh launch script.
• At the time of submission, the vLLM Kimi-K2.5 recipes page does not have a B300-specific recipe, so this reuses the existing Kimi-K2.5 INT4 B200 vLLM recipe as-is until B300-specific tuning is available.
• Image: vllm/vllm-openai:v0.20.0-cu130 — the original drafts (Add B300 config: kimi-k2.5-int4-vllm #1057, Add B300 config: kimi-k2.5-int4-vllm (vLLM 0.20.0 + TP=4/EP=1 sweep) #1071) carried the v0.19.0 placeholder while we waited; vLLM 0.20.0 has now shipped.
• Search-space per (ISL, OSL):
  • { tp: 8, conc-start: 4, conc-end: 64 }
  • { tp: 4, ep: 1, conc-start: 4, conc-end: 64 } (new — covers the lower-TP / expert-parallel variant on the same B300 nodes)
• perf-changelog.yaml entry added at the top, pointing at the original tracking PR Add B300 config: kimi-k2.5-int4-vllm #1057.

Test plan

• CI config validation passes
• Run kimik2.5-int4-b300-vllm single-node benchmark on a B300 node and confirm the v0.20.0 image starts, both TP=8 and TP=4/EP=1 sweeps complete, and result files are produced for both ISL/OSL configs.

🤖 Generated with Claude Code

[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://cookbook.sglang.io/
• https://docs.vllm.ai/projects/recipes/en/latest/

If it is not, please create a PR first before we can merge your 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/sgl-cookbook

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]

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://cookbook.sglang.io/
• https://docs.vllm.ai/projects/recipes/en/latest/

If it is not, please create a PR first before we can merge your 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/sgl-cookbook

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=25267828360
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=25267828360