WorkBuddy custom model not taking effect after saving—did I miss a step?

I followed that four-step process for adding a custom model in WorkBuddy, and I got stuck on the last step. I still can't figure out if I filled something in wrong or if it's just like that.

After my night shift, I came across a tutorial on how to hook a third-party API into WorkBuddy, saying it's done in four steps. I've been working on a quick drug reference table for my department, and I've always used the default model. When I saw I could switch, my hands got itchy and I wanted to try it. I've used it for two months, and this is the first time I've dug this deep.

Go into settings, models on the left, plus sign in the top right. Provider: custom, then I filled in the endpoint URL, API Key, and model name. The key was created in the console, starts with sk-. Save, it says success.

Go back to the chat page and send a message, but it still replies with the original one, even the tone hasn't changed. I guessed I missed the switching step. The tutorial said after saving, click the model name at the bottom right of the input box and select it from the custom list. I clicked, but I didn't seem to see the one I added in that list—or maybe I was looking in the wrong place.

If you don't switch after saving, it keeps using the old model—I only saw this when I checked the error troubleshooting table, and then it clicked.

In between, I also tried checking the thinking mode, and sending it immediately gave a 400 error. After turning it off, it was fine. The tutorial mentioned that the checkbox doesn't give the model new abilities, so don't enable capabilities you haven't confirmed are supported. I didn't run into a 403 error, so it seems the key part passed.

So I want to ask, has anyone successfully run a custom model? Does that list take a while to refresh, or did I fill in something wrong?