The latest articles on DEV Community by Stephen Chukwunonso (@stephen_chukwunonso_d170e). https://dev.to/stephen_chukwunonso_d170e https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4189574%2F48179b44-3b9b-4f1b-bfb4-1000352e271a.png https://dev.to/stephen_chukwunonso_d170e en Stephen Chukwunonso Fri, 16 Oct 2026 21:54:05 +0000 https://dev.to/stephen_chukwunonso_d170e/how-do-you-keep-prompts-tools-and-tests-working-when-you-switch-models-5dd https://dev.to/stephen_chukwunonso_d170e/how-do-you-keep-prompts-tools-and-tests-working-when-you-switch-models-5dd <p>I've been thinking about what happens to an AI agent when the model underneath it changes. The code doesn't change, but the behavior can. A prompt that worked well on one model may give different answers on another, and a tool call that was reliable may start failing quietly.<br> How are you handling this today? Some questions I'm curious about:<br> –When you upgraded or switched models, what broke first?<br> –Did you have tests for agent behavior, or did you find out from users?<br> –Are your prompts, tool definitions, and test cases stored together, or in different places?<br> –Did you build any manual process to catch these changes, and what does it look like?<br> –What's the last model change that surprised you?<br> –How do you know when an agent has gotten worse?<br> –What do you do before you upgrade a model?<br> –What tool do you wish existed for this?<br> If you've solved this, I'd like to hear how. If you haven't, I'd like to hear that too. Share what you've run into in the comments.</p> agents ai discuss llm