GPT-4 Turbo stopped working suddenly

You're sending requests to GPT-4 Turbo and getting errors, or the model just isn't available anymore in your dropdown menu. This hits both API users and ChatGPT Plus subscribers, often without warning.

GPT-4 Turbo stopped working suddenly

What's actually happening

OpenAI retired the original `gpt-4-turbo-preview` model name on December 11, 2024, and replaced it with the stable `gpt-4-turbo-2024-04-09` identifier. If your code or application still references the old preview version, your API calls now return a 404 "model not found" error. You'll see something like: `Error: The model 'gpt-4-turbo-preview' does not exist or you do not have access to it`.

For ChatGPT web users, GPT-4 Turbo might disappear from your model selector if your Plus subscription lapsed, your payment failed, or you hit your usage cap for the billing period. OpenAI enforces rate limits even on paid plans — when you exceed them, the platform temporarily downgrades you to GPT-3.5 or GPT-4o mini until your limits reset.

Another common cause: you're still subscribed, but OpenAI switched your default model to GPT-4o during a platform update. Check the dropdown at the top of your chat window — if it says "GPT-4o" instead of "GPT-4 Turbo", you've been moved automatically. This happened to thousands of users during the GPT-4o rollout in May 2024, when OpenAI made the newer model the default for all Plus subscribers.

How to fix it

For API users:

Open your codebase and search for `gpt-4-turbo-preview` or any model identifier containing "preview"
Replace it with `gpt-4-turbo-2024-04-09` — this is the stable, permanent identifier
Test your request with a simple curl command:

```

curl https://api.openai.com/v1/chat/completions \

-H "Authorization: Bearer YOUR_API_KEY" \

-d '{"model":"gpt-4-turbo-2024-04-09","messages":[{"role":"user","content":"test"}]}'

```

If you still get a 404, check your API key permissions at platform.openai.com/api-keys — make sure the key has access to GPT-4 models
Verify your organization hasn't restricted GPT-4 access in the Usage policies section under Settings

For ChatGPT web users:

Go to chatgpt.com and click your profile picture in the bottom left
Select "My plan" — confirm your Plus subscription shows "Active" with a next billing date
If it says "Cancelled" or shows a failed payment, update your payment method under "Manage my subscription"
Return to the chat interface and click the model dropdown at the top of the screen
Look for "GPT-4 Turbo" in the list — if it's there, select it manually
If you see "You've reached the current usage cap" below the dropdown, you've hit your limit for GPT-4 messages this billing cycle (typically 40 messages every 3 hours for Plus)
Wait for the timer to reset, or switch to GPT-4o which has higher limits

If GPT-4 Turbo isn't in your dropdown at all and your subscription is active, sign out completely, clear your browser cache, and sign back in. This forces the platform to refresh your account permissions.

If that doesn't work

Contact OpenAI through the help centre at help.openai.com — click "Messages" in the bottom right corner. Include your account email, the exact error message (screenshot it), and your subscription status. For API issues, include your organization ID from platform.openai.com/settings/organization and the specific model name you're trying to use.

Most access issues get resolved within 24 hours if it's a billing problem. Model deprecation questions usually get a response within 2-3 business days. If you need to understand other behaviour changes, the platform status page at status.openai.com shows current incidents and model availability.

Questions people actually ask

Q: Will my old GPT-4 Turbo code break permanently?

A: The preview versions are gone for good. Update to `gpt-4-turbo-2024-04-09` immediately — it won't auto-redirect.

Q: Why can't I switch back to GPT-4 Turbo from GPT-4o?

A: GPT-4o is now the default for most tasks. GPT-4 Turbo still appears in the dropdown if your account supports it, but OpenAI prioritizes the newer model for speed and cost efficiency.

Q: Does GPT-4 Turbo cost more than GPT-4o?

A: For API users, GPT-4 Turbo costs $10 per million input tokens versus $2.50 for GPT-4o. ChatGPT Plus subscribers pay the same $20/month regardless of which model they use, but GPT-4o has higher usage caps.

What to remember

Replace `gpt-4-turbo-preview` with `gpt-4-turbo-2024-04-09` in all API code immediately
ChatGPT Plus usage caps reset every 3 hours for GPT-4 models — check the timer in your model dropdown
Failed payments automatically revoke access to GPT-4 models within 24 hours
OpenAI now defaults most users to GPT-4o — manually select GPT-4 Turbo from the dropdown if you need it
Check platform.openai.com/usage for detailed logs of your API requests and rate limit status

Related help

---

*openai-support.com is an independent resource, not affiliated with OpenAI Inc.*