Using Compare Mode

Compare Mode sends the same prompt to multiple AI models at once and shows their responses side by side. Use it to check accuracy, reasoning, tone, and output quality before choosing a response.

Learn more about Compare Mode →

How to Compare AI Models Side by Side

• Open a new chat.
• Enable the Compare toggle in the chat input area.
• Select the models you want to include. Up to 6 models can be compared at a time.
• Enter your prompt and send it.
• Review the responses side by side.
• Starting with two or three models makes the responses easier to review closely.

When to Use Compare Mode

• Starting a new type of task and unsure which model handles it best
• Checking a fact or a piece of reasoning against more than one model
• Choosing a default model for recurring work, such as product descriptions or code review
• Reviewing multiple perspectives before making a decision

For quick questions, a single model is usually enough.

How to Write a Prompt for Comparing Models

A comparison is most useful when the prompt reflects real work.

• Use a real task: Paste the actual email, code, or brief instead of a sample question.
• Define the output: State length, format, and tone. For example, "Under 150 words, friendly, ends with a clear next step."
• Ask for something checkable: Facts, code that runs, or calculations are easier to evaluate than open-ended creative requests.
• Keep the context equal: Avoid sending follow-up messages to only one model if you want a fair comparison.

How to Evaluate AI Model Responses

Review each response against the same criteria:

• Accuracy: Are the facts correct?
• Instruction following: Does it match the requested length, format, and tone?
• Reasoning: Does the logic hold up for analysis or code?
• Tone: Is it ready to send or publish?
• Editing effort: Which response needs the fewest changes?

Example Comparison Prompts by Task

• Writing: "Rewrite this paragraph for small business owners. Keep it under 80 words and avoid jargon."
• Coding: "This function throws an error on empty input. Fix it and explain the change in two sentences."
• Research: "Summarize the main arguments for and against a four-day work week, and note where the evidence is weak."
• Data: "Here is a table of monthly sales. What is the trend, and what looks unusual?"

Why AI Models Give Different Answers

AI models are trained on different data, with different architectures and goals. Some focus on reasoning, others on speed, creativity, or technical accuracy. As a result, the same prompt can produce different structure, length, and conclusions.

When models disagree on a fact, verify it before using the response.

Compare Mode vs. Switching Models

The model selector lets you chat with one model at a time and switch whenever needed. Compare Mode sends one prompt to several models at once, so responses can be reviewed together in a single view. See Using Aymo AI for details on switching models.

Tips for Getting More from Compare Mode

• Note which model performs best for each type of task.
• Re-test occasionally, since new models are added regularly.
• Continue the conversation with the model that gave the best response.