[thomhurst]

The homepage race defaults to the JIT view, so the AOT numbers were only visible after clicking the Native AOT toggle, and nothing compared AOT against JIT.

• Adds a TUnit (AOT) lane at the bottom of the JIT view, labelled with how much faster it is than TUnit on JIT (e.g. 3.6× for async, ~24× for data-driven in the latest run).
• The headline "N× faster than the next fastest" figure is unchanged: still JIT against JIT.
• The Native AOT toggle (TUnit AOT vs xUnit v3 AOT) is unchanged.
• Deletes the unused BenchmarkHighlight component, which nothing imported and which had a hardcoded "10x faster with Native AOT" claim, and updates BENCHMARKS_IMPLEMENTATION.md to match.

Checked with tsc and a full docusaurus build; the rendered homepage has the new lane.

[chatgpt-codex-connector]

Codex Review Summary: ✅ Completed

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

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-11T18:32:15.492371Z	75bebdd	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.

[coderabbitai]

Caution

Review failed

The pull request is closed.

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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

[github-actions]

Review

The change is small and does what the description says. The AOT lane is added to the JIT view, the headline "N× faster" figure is still computed from JIT lanes only, and the dead BenchmarkHighlight component and its docs entry are removed. I only read the diff and did not run tsc or docusaurus build.

Minor suggestions:

• Label when AOT is slower. The label (tunit.ms / lane.ms).toFixed(1)× faster than JIT would read "0.9× faster" if a run ever had AOT slower than JIT. Branch on the ratio, or fall back to "× of JIT time", so the homepage can't show a nonsensical claim after a bad benchmark run.
• Index arithmetic. index >= lanes.length is repeated three times to identify the AOT lane. Rendering lanes and aotLanes as separate arrays, or tagging each lane with kind: 'aot', would be clearer and less fragile if more reference lanes are added.
• Type of aotLane. runtimes is untyped, so aotLane is optional only implicitly. A small Runtime type would make that explicit.

None of these block merging.

[greptile-apps]

Confidence Score: 4/5

[Low impact] The changes affect the homepage benchmark display and its implementation notes.

The PR appears safe to merge, with a non-blocking improvement to the AOT comparison label.

Findings

1.  Slowdowns are labelled faster ▶

Summary

Adds a TUnit (AOT) reference lane to the homepage’s JIT race.

• The JIT race shows TUnit’s Native AOT result beside its JIT results.
• The unused benchmark card component and its fixed AOT claim are removed.

_{Reviews (1) · Last reviewed commit: "docs: show TUnit Native AOT in the homep..." · Reviewed by Greptile}

[greptile-apps]

Slowdowns are labelled faster

The new AOT label always says “faster than JIT,” even when lane.ms exceeds tunit.ms. If a future benchmark update reports slower AOT results, visitors would see something like “0.8× faster than JIT.” Handle slower and equal results explicitly, as the headline already does, or use a neutral ratio label.