Fix model not found error chat in ChatGPT

If you see a "Model not Found" error when opening Chat, it usually means your account or organization does not have access to the model the Chat UI is trying to use, the model name in the UI differs from what your API key can call, or your browser is showing a stale UI. This guide walks through quick checks, an API fallback you can run now, and tool specific notes for n8n, Make.com and Zapier.

Why you get the model not found error chat

There are three common root causes:

Account or organization restrictions: your API key or account is not permitted to use the model the Chat UI selected. OpenAI can restrict certain models by plan or organization.
Model name mismatch: the UI asks for a model id that no longer exists or is renamed, so the backend returns the text "Model not Found".
Browser or UI cache problem: a stale client-side cache or corrupted Local Storage prevents the Chat UI from seeing available models or it keeps requesting an old model id.

How to fix model not found error chat: step-by-step

Confirm the exact error text. If you see "Model not Found" or "This response couldn't load" note it exactly. Capture the console network request id where possible for support.
Quick UI checks. Sign out of chat.openai.com, open an incognito/private window and sign back in. If models appear, the problem is browser cache or Local Storage.
Clear Chat Local Storage. In Chrome/Edge: open DevTools (F12), go to Application, under Local Storage select https://chat.openai.com, delete keys for model or full storage, then hard reload (Ctrl+Shift+R). On desktop app sign out and quit the app before restarting.
Verify account and org model availability. Open the OpenAI Dashboard, go to your profile icon and check Organization settings and billing. Make sure your organization has the model available and no billing holds. If you have multiple organizations, ensure the correct org is selected in the top-right of the dashboard.
Run an API model list to confirm what your key can see. Use the reproducible curl below. If the model id the UI expects is missing, you will see the same problem via API.
```
curl https://api.openai.com/v1/models \
  -H "Authorization: Bearer $OPENAI_API_KEY"
```
Search the list for the model id the Chat UI requested. If you do see common models like gpt-3.5-turbo but not newer ones, your account or plan may not include the missing model.
Test a direct chat completion call. This verifies the API path works independent of the UI. Example with a common available model:
```
curl https://api.openai.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{"model":"gpt-3.5-turbo","messages":[{"role":"user","content":"Say hello"}],"max_tokens":50}'
```
If this returns a normal response, your API key and network are fine and the issue is likely client-side or organization-specific.
If the API call also says model not found, contact your admin or OpenAI support. Include the exact request, timestamp, and any response headers or request id from the curl output. You can link to the OpenAI forum threads if you need to show others the problem, for example this thread about "Model not Found" in Chat and a report of missing models on desktop and web.

Tool specific notes: n8n, Make.com and Zapier

When your automations stop because of missing models, check where the call is made and which credentials and model name are used.

n8n: If you use the built in OpenAI node, open the node and confirm the Model field exactly matches a model id your key can access. For custom calls use the HTTP Request node and run the same curl payload as a POST. If your n8n workflow fails with "Model not Found", go to Credentials in n8n, re-select your OpenAI API Key, and test the credential. If you use JavaScript inside a workflow, remember the custom JS node is named Code, not Function.
Make.com: Edit the scenario step that calls OpenAI. If you use the OpenAI module, re-authorize the connection and verify the Model dropdown shows the model you expect. For HTTP requests use the HTTP module and paste the JSON body shown above. A common fix is to re-save the scenario after re-authorizing so Make refreshes available models.
Zapier: Open the Zap and check the OpenAI action step. If the step shows an error, reconnect the OpenAI account, then test the step. Zapier also supports Webhooks by Zapier for POSTs; use that to run the same test JSON if you need to bypass the native app temporarily.

Browser and UI cache checks that work in Nigeria and elsewhere

Browsers can hold old model IDs in Local Storage which results in the UI repeatedly requesting a model the account no longer has. Do this sequence:

Open a private window and sign into chat.openai.com. If things work there, clear the normal profile Local Storage for chat.openai.com as described earlier.
Clear cookies for chat.openai.com, then reload. If you use the desktop app, sign out and choose Quit app from the menu, then re-open.
If you manage multiple accounts or orgs, make sure you select the right organization from the profile menu when logged in. You will get "Model not Found" if the UI tries to call a model from an org you are not currently in.

How to stop this happening again

Use explicit model names in automations, and keep them in configuration, not hard coded in UI templates. That helps you rotate models without the UI picking an unsupported id.
Keep a small troubleshooting script or saved curl commands to run the model list and a chat call. Put them in a secure notes file so you can run them on any machine.
For teams, assign one admin to monitor billing and plan limits. Missing billing or plan changes are common causes of model access removal.
When deploying to production workflows in n8n, Make.com, or Zapier, include a catch branch that logs API errors and alerts you with the error text, so you see "Model not Found" quickly and can switch to a fallback model.

When to get help

If after clearing Local Storage and confirming your API key cannot see the model via GET /v1/models, open a support request with OpenAI including the curl request and the full response. Community threads can help if the issue is widespread, for example users reporting missing models on desktop and web or the forum thread where people posted "Model not Found" in Chat. If you manage automation for a business and the problem is blocking operations, escalate by providing the OpenAI request id from your failing requests and timestamps.

For more on Chat UI response problems see our related post ChatGPT not showing responses, step-by-step fix.

If you want hands on practice building resilient automations that include API fallbacks and alerts, learn with Automate Naija.

Frequently asked questions

What does "Model not Found" mean in ChatGPT?

"Model not Found" means the chat client tried to request a model id the server could not find for your account. That can be due to account organization restrictions, a renamed or removed model id, or a stale browser UI requesting an old id. Check account access and run an API model list to confirm.

How do I check which models my API key can use?

Run a model list with curl: curl https://API.OpenAI.com/v1/models -H "Authorization: Bearer $OPENAI_API_KEY". The response lists model ids your key can access. Search that output for the model id the UI or your automation is trying to use.

Why do models disappear from the Chat UI on desktop and web?

Models can be hidden from your org by plan changes, billing holds, or admin policy. Also the browser may have cached an old model id in Local Storage so the UI requests a missing model. Try an incognito session, clear Local Storage for chat.OpenAI.com, and verify your organization and billing status.

My automation in n8n/Make/Zapier fails with the same error, what should I do?

Check the exact model string used in the node or module. In n8n verify the OpenAI credential and test with an HTTP Request node. In Make.com re-authorize the OpenAI module and resave the scenario. In Zapier reconnect the OpenAI account or use Webhooks by Zapier to run the same POST payload to confirm API access.

Build it with us. The community is free and the build track teaches this kind of automation week by week: see pricing or join the community.