Loading page content
Put two prompt variants side by side and compare them on clarity and specificity scores, five quality signals, and estimated token cost, with a headline winner.
Compare two versions of a prompt side by side. See token counts, estimated costs, clarity scores, and which prompt includes best practices like role definition, output format, constraints, and examples.
Prompt A Analysis
Prompt B Analysis
Cost estimate based on GPT-4.1 pricing ($2.00/1M input tokens). A higher score suggests the prompt follows more prompt engineering best practices.
This scores structure, not outcomes. It tells you which prompt is clearer, more specific, and cheaper, which correlates with better results but does not replace running both against real inputs and grading the outputs. For anything going to production, build an eval set; the Prompt Eval Suite exists for that.
Where it earns its keep is the cost axis. A prompt variant that scores marginally better but costs 40% more per call is a bad trade at volume, and that trade-off is invisible when you are comparing outputs by eye in a playground.