Skip to content
runlog/
Developers / Private comparisons

Compare through
the advanced workflow.

For an on-demand private comparison with expected cost, maximum spend, and actual cost, open Studio. No API key is needed there. Use this advanced Benchmark workflow for API access, explicit suites, and integrations.

Run a comparison

For this advanced workflow, create an API key in Account. Paste it below; it stays on this page and is not included in drafts or reports. Model execution sends your prompt to configured providers.

Loading available models…

exact:text, contains:text, json_valid, length:10..500, number:/score:0..10, or regex:^OK, one per line. Check syntax and JSON Schema examples. Saved multiline checks use a JSON array to preserve their exact text.

Save to Workspace (optional)

Use “Save draft in Workspace” below, save a suite, then choose “Compare this suite” to fill these IDs automatically. You can also enter IDs from Workspace. The key must belong to the same account and the suite must match this prompt and its checks.

No comparison has been run.

Continue the loop

What happens next?

Use the comparison to choose a reviewed baseline. Use Workspace for private signed-in storage of project suites and retained reports. Return to Studio for the next on-demand comparison. Benchmark reports alone are not Ship-Check release evidence. Use your own measured tool trace to set up Ship-Check, then add local runtime limits with Loop-Guard.