Compare prompts side-by-side

A common development workflow is compare multiple prompts side-by-side. The prompt playground allows users to directly compare up to three prompts at a time. This makes it easy to review changes and determine whether a modified prompt performs better than the basline and alternatives.

Create a new prompt for side-by-side comparison.

Duplicate Prompt A and then switch the model from gpt-3.5-turbo to gpt-4o for a direct comparison.

Ask Copilot to optimize the prompt to reduce the likelihood of the LLM generating hallucinated responses.

Click on the row to compare all 3 outputs side-by-side in full screen.