When using ChatGPT, what most affects the experience is often not the features, but “messages won’t send,” “the model can’t be used,” and “it keeps asking for verification.” This article breaks down common symptoms by scenario and provides a troubleshooting order you can follow immediately to avoid detours as much as possible.
1. ChatGPT can’t send messages or shows a request error
First check whether it’s a temporary glitch: refresh the page, log out and back into ChatGPT, then retry the same message. If it keeps showing errors like “Something went wrong,” prioritize switching network environments or switching to mobile data to rule out local network blocking or instability.
Next, check the browser: disable ad blockers and script-related extensions, then open ChatGPT in an incognito/private window. If it still doesn’t work, clear the site cache and cookies (only for ChatGPT-related sites is enough). Many cases of “can’t send messages” are actually due to an expired old session token.
2. ChatGPT models unavailable, empty dropdown list, or switching fails
Model list issues are usually caused by account status or incomplete loading of page resources. First confirm whether ChatGPT can normally load the sidebar and chat history; if page elements are missing, try a hard refresh (Windows: Ctrl+F5 / Mac: Cmd+Shift+R).
If the interface is visible but model switching keeps failing, try starting a new conversation instead of repeatedly switching within an old one. When a conversation context is large, ChatGPT is more likely to get stuck during model switching and context restoration.


