How to Get GPT-4o Back: ChatGPT and API Options

Short answer: You can’t restore GPT-4o in ChatGPT. OpenAI retired it from regular ChatGPT chats on February 13, 2026. Business, Enterprise, and Edu customers kept access inside Custom GPTs until April 3, 2026; OpenAI says it was fully retired across all ChatGPT plans after that date. OpenAI’s API documentation still lists GPT-4o, but marks it deprecated—so API access is a separate, potentially temporary route, not a way to bring it back in ChatGPT. (help.openai.com)

Status checked November 12, 2026. This guide covers OpenAI’s own ChatGPT and API; it does not establish GPT-4o availability through another provider.

If you mean GPT-4o in ChatGPT

There is no current ChatGPT setting, plan upgrade, or model-picker step that restores GPT-4o. OpenAI’s retirement notice says the model is no longer available in ChatGPT. The April 3, 2026 cutoff applied to Business, Enterprise, and Edu customers using GPT-4o within Custom GPTs; after that, OpenAI says GPT-4o was fully retired across all plans. If you can’t see it in ChatGPT, changing a setting or subscription is not a documented fix. (help.openai.com)

OpenAI says ChatGPT conversations and projects that used retired models default to their GPT-5.3 Instant and GPT-5.4 Thinking or Pro equivalents. ChatGPT Voice and Images are described as unchanged, but they are separate experiences—not a way to select the retired text GPT-4o model. (help.openai.com)

If you mean GPT-4o through a production API

ChatGPT retirement does not itself mean API retirement: OpenAI says GPT-4o remains available through the API. Its model page currently labels GPT-4o Deprecated and lists the gpt-4o model ID and the Responses API as a supported endpoint. Confirm that your API project can access the model before depending on it. (help.openai.com)

Make a minimal OpenAI API request

This example uses OpenAI’s Python SDK and Responses API. It requires OpenAI API access and a project API key. ChatGPT and API billing are separate; a ChatGPT subscription does not, by itself, pay for API usage. (developers.openai.com)

1. Create an API key in the OpenAI Platform and store it securely.
2. Install the SDK and set your key as an environment variable:

pip install openai
export OPENAI_API_KEY="your_api_key_here"

3. Save this as test_gpt4o.py and run python test_gpt4o.py:

from openai import OpenAI

client = OpenAI()
response = client.responses.create(
    model="gpt-4o",
    input="Reply with: GPT-4o API is responding."
)
print(response.output_text)

If the model is available to your project and the request succeeds, the script prints the model’s reply. The GPT-4o model page lists the Responses endpoint and the gpt-4o identifier. (developers.openai.com)

Check model access and shutdown information

OpenAI documents checking the Models API for model availability and an optional shutdown_date. For example, use your API key to list models and inspect the GPT-4o entry:

curl https://api.openai.com/v1/models \
  -H "Authorization: Bearer $OPENAI_API_KEY" | jq '.data[] | select(.id == "gpt-4o") | {id, shutdown_date}'

A returned gpt-4o entry indicates that the model appears in that API organization’s model list; it does not guarantee uninterrupted future access. OpenAI says shutdown_date is optional and can be null when no shutdown date has been announced. The current model page marks GPT-4o deprecated but does not state a specific shutdown date in the information cited here. (developers.openai.com)

Cost and production caveat

OpenAI lists standard GPT-4o text pricing at $2.50 per 1 million input tokens, $1.25 per 1 million cached input tokens, and $10 per 1 million output tokens. For example, 100,000 uncached input tokens and 20,000 output tokens would cost about $0.45: $0.25 for input plus $0.20 for output. Actual charges depend on usage and any other applicable features or charges. (developers.openai.com)

Because GPT-4o is deprecated, test a migration path before making it a production dependency. Check the model listing and OpenAI’s deprecation documentation for updates; an absent or null shutdown_date is not a guarantee that the model will remain available indefinitely. (developers.openai.com)

About TokenForge Cloud: The API example above is for OpenAI’s direct API only. The sources for this article do not establish whether TokenForge Cloud currently offers GPT-4o, or document a TokenForge model ID, endpoint, credentials, or price. Confirm those details in the provider’s own documentation rather than assuming OpenAI’s setup will work there.

If GPT-4o is missing or the API request fails

• Missing in ChatGPT: GPT-4o has been retired there, so switching plans or changing settings is not a documented way to restore it. Choose a model currently offered in ChatGPT instead. (help.openai.com)
• API request fails: Check that your API key is set correctly, the request uses gpt-4o with a supported endpoint, and the model appears in your API organization’s model list. Check API billing separately from ChatGPT billing. These are access and setup checks; a failed request by itself does not establish that the model has been shut down. (developers.openai.com)
• Still blocked: Check OpenAI’s API status and support resources for service or account issues. If OpenAI has shut down the model for your project, a ChatGPT setting change will not restore it; plan a migration. (developers.openai.com)

Bottom line

There is no supported way to get text GPT-4o back in ChatGPT. OpenAI’s API documentation still lists gpt-4o, but marks it deprecated. Check whether your API organization can access it, review any announced shutdown date, account for separate API billing, and prepare a migration plan before using it in production.

Sources

Retiring GPT-4o and other ChatGPT models — OpenAI Help Center

GPT-4o Model — OpenAI API

Deprecations — OpenAI API

Managing billing for ChatGPT and the API platform — OpenAI Help Center