ChatGPT FAQ: How to Handle Frozen Chats, Missing History, and Unavailable Models

When using ChatGPT, the most infuriating thing isn’t not knowing how to ask questions—it’s when it suddenly freezes, won’t send, or your history seems to “disappear.” Below is a ChatGPT FAQ organized by the most common scenarios, using actionable steps to help pinpoint the issue as clearly as possible. You can troubleshoot in order, and in most cases things return to normal within a few minutes.

What to do when the chat doesn’t respond, keeps spinning, or fails to send

If ChatGPT keeps spinning or becomes unresponsive, first switch to a different network environment (for example, from a corporate network to a mobile hotspot) to rule out unstable connections caused by a proxy, DNS issues, or network blocking. Then refresh the page and log back into your account—many “freezes” are actually due to an abnormal session state.

If it still fails, try disabling browser extensions (ad blockers, script managers, and translation extensions are the most common culprits), then clear your cache and cookies. Finally, check OpenAI’s status page to see whether the service is experiencing fluctuations: if there’s an official outage, there’s nothing you can do but wait for recovery.

Chat history is missing: is it truly gone or just a display issue?

First, confirm whether you switched accounts or login methods (email login vs. Google/Apple login—if they don’t match, it will feel like you “became a different person”). Chat history is tightly bound to the account; if you log into the wrong account, the left-side list will look empty.

Next, check whether you turned off the “chat history” related option in settings: after it’s turned off, new conversations may not appear in the history list, but that doesn’t mean there’s an account problem. If the list simply won’t load, try a hard refresh, switch browsers, or log in using an incognito window—often it’s a local cache issue causing a display glitch.

Model unavailable, can’t switch, or “insufficient permissions”

When ChatGPT says a model is unavailable, switching fails, or you lack permission, first rule out “temporary capacity limits” and “server-side fluctuations.” You can log out and back in, or open a new chat window and try again, to avoid an old conversation staying stuck in an abnormal state.

If you joined a team/workspace under the same account, note that conversations may appear under different spaces, and permissions can differ after switching workspaces. Another common cause is the browser blocking site scripts: add chat.openai.com to the allowlist, which usually fixes issues where the model dropdown doesn’t load completely.

Account locked, repeated verification, or “suspicious activity”

If ChatGPT repeatedly asks for verification or flags suspicious activity, it’s usually related to frequent IP changes, shared networks, or unstable proxy nodes. Use a stable network environment for a while, avoid rapidly switching logins across multiple devices in a short time, and try not to use unknown proxies.

If your account is locked or you can’t pass verification, follow the on-page instructions to complete email/phone verification, and check your spam folder in case verification emails were mistakenly filtered. If you still can’t recover access, you’ll have to submit an appeal through official support channels; before submitting, prepare your login email, screenshots of the issue, and the approximate time it occurred—this makes it easier to locate and resolve.