Titikey
HomeTips & TricksChatGPTChatGPT FAQ: Fixes for Missing Chat History and Unavailable Models

ChatGPT FAQ: Fixes for Missing Chat History and Unavailable Models

2/22/2026
ChatGPT

When using ChatGPT, the most infuriating thing is often not the quality of the answers, but conversations suddenly disappearing, models not opening, or messages failing to send. Below, these high-frequency ChatGPT issues are broken down by scenario and explained clearly, with a practical troubleshooting order. Follow the steps and you can basically pinpoint the cause.

Chat history missing: First confirm whether it was “not saved”

When ChatGPT chat history disappears, the most common reason is logging into the wrong account or switching the login method (email/Google/Apple). It may look like the same email, but it’s actually a different account. First, check the account avatar and email in the bottom-left of ChatGPT, then log in again in an incognito/private browser window to verify.

Next, check the chat history option in Settings: if you enabled Temporary Chats or turned off chat history saving, new conversations simply won’t appear in History. On the ChatGPT web app, it’s also recommended to refresh the sidebar and try logging in on another device to see whether it’s just that the local cache hasn’t updated.

Model unavailable or unable to switch: Usually network and session state

If ChatGPT shows “Model unavailable” or “Unable to load model list,” first rule out network blocking: switch networks (mobile data/other Wi‑Fi) and disable any proxies, ad-blocking scripts, or privacy extensions that might interfere. Often the ChatGPT homepage loads, but the model-list API is blocked, which shows up as an empty dropdown or a grayed-out button.

If you can’t switch models only within a specific conversation, start a new ChatGPT chat and try again. This quickly tells you whether the session is stuck or it’s a global issue. You can also check the OpenAI status page for service disruptions; when there’s an incident, repeatedly refreshing only makes it easier to trigger rate limits.

Send failed, request error: Check item by item from input content to browser environment

If ChatGPT indicates a send failure or request error, first copy the content in the input box locally, delete it, and send it in smaller segments to avoid a single overly long message or unusual characters causing the request to be rejected. If you pasted large blocks of code, tables, or mixed-format text, it’s recommended to paste as plain text first, then gradually restore formatting.

On the browser side, clear the ChatGPT site cache and cookies, or test with a different browser—common causes are extension conflicts or corrupted cached data. In a corporate network environment, if ChatGPT errors frequently, try changing DNS or switching to a mobile hotspot to more quickly determine whether a gateway is blocking it.

Account issues and security prompts: Don’t brute-force refresh—reduce risk first

If ChatGPT shows security verification, unusual activity prompts, or logs you out frequently in a short period, it’s usually related to logging in from different locations, frequently changing IPs, or automation plugins. First disable auto-refresh/script-type extensions, keep the network stable for a while, then log back into ChatGPT—the success rate will improve noticeably.

If you suspect someone else has logged into your account, change your password immediately and review logged-in devices; also enable stricter email security settings to prevent the account from being stolen again. If you encounter restriction prompts you can’t resolve on your own, submit a ticket via the Help entry on the ChatGPT page and include an error screenshot and the approximate time it occurred to speed up diagnosis.

HomeShopOrders