A better prompt. A clearer result.
Define your goal, test across models, and keep the best version.
Evaluation
ReadyAll selected models evaluate the same prompt.
Improvement from your starting prompt will appear here.
Run an optimization to see feedback for each model.
Evidence & scoring method
Clarity: Task, audience, and expected response are clear.
Goal alignment: Instructions support the objective and next action.
Constraints: Tone, length, and missing-information rules are explicit.
Clarity × 30% + goal alignment × 40% + constraints × 30%. Scores use one decimal; weighted totals use two decimals for comparisons.
The target applies to each model total, not every individual criterion. The headline uses the lowest total. Demo scores are illustrative and do not measure real output quality.
Your improved prompt
Best version appears here after evaluation.
ORIGINAL PROMPT
OPTIMIZED PROMPT
Changes are summarized after evaluation.
Iteration history · no runs yet
| Version | Lowest / 10 | Revision |
|---|
Illustrative progression for this demo. Real evaluations may improve or regress.