← Evals & Launch Decisions / 16
hardNewCursorDatabricksStripe

Our full eval suite takes six hours and costs a few hundred dollars per run, so engineers have stopped running it. What do you do?

Try it in the practice workspaceDraft your answer before comparing it with the rubric.

An eval nobody runs is worth nothing. This question tests whether you can do the cost arithmetic and redesign the workflow, rather than ordering people to comply.

Updated Nov 2026 · Original practice questions with worked answers and teaching rubrics. Our editorial policy explains the sources, review process and limits.

Follow-up roundsAnswer first, then revise for each changed constraint
  1. Round 2. Schema checks would reject 10% of outputs before judging, and a judge at a tenth of the price matches the expensive one on two of the four criteria, validated against human labels.
  2. Round 3. A model upgrade lands. Against the current model, the full suite shows 41 newly failing and 63 newly passing cases out of 2,000, and the release is due tomorrow.

An eval nobody runs is worth nothing. This question tests whether you can do the cost arithmetic and redesign the workflow, rather than ordering people to comply.

This reference answer is in Premium. Free signup opens other answers.

9 answers per topic instead of 5, and your progress kept · no cardor unlock all 420 remaining answers
UP NEXT ON YOUR JOURNEY
NOTES & DISCUSSION · 0

No comments yet. How would you have answered?