No—not in ChatGPT. OpenAI retired GPT-4o from ChatGPT on February 13, 2026. Business, Enterprise, and Edu customers retained it in Custom GPTs until April 3, 2026, when it was retired across all plans. OpenAI says GPT-4o remains available through the API. This information was checked on November 9, 2026. (OpenAI Help Center)
ChatGPT Free is not free GPT-4o
ChatGPT has a free plan, but GPT-4o is no longer one of its ChatGPT models. OpenAI’s current Free Tier FAQ names GPT-5.6 Luna and notes that the default model and available limits can change. OpenAI also says free users have unlimited text chats; tools such as file uploads, image creation, and data analysis have separate limits. (OpenAI Free Tier FAQ; OpenAI announcement)
GPT-4o’s retirement from ChatGPT does not mean it is free through the API. OpenAI says it remains available there, but its API model documentation marks GPT-4o deprecated. Check the current model directory before building around it. (OpenAI Help Center; GPT-4o API model page)
Would ChatGPT Plus bring GPT-4o back?
No. OpenAI lists ChatGPT Plus at $20 per month, billed monthly. Plus provides higher model limits and broader model access than Free, but GPT-4o’s retirement applies across ChatGPT plans. A Plus subscription does not include general API usage; API usage is billed separately. (OpenAI Help Center)
What does GPT-4o API access cost?
OpenAI’s GPT-4o model page lists text-token rates of $2.50 per 1 million input tokens, $1.25 per 1 million cached input tokens, and $10 per 1 million output tokens. The page displays dollar signs but does not spell out a currency code or specify the processing mode in the rate row. For the illustration below, the dollar amounts are treated as U.S. dollars; confirm the currency and applicable rate in your API billing details, especially if using Batch or another processing option. These are token-based API charges, not a monthly GPT-4o subscription. (developers.openai.com)
For a simple text-only estimate using those listed rates:
Estimated cost = (uncached input tokens × $2.50 + cached input tokens × $1.25 + output tokens × $10) ÷ 1,000,000
Example, not a usage test: 1 million uncached input tokens plus 100,000 output tokens would cost an estimated $3.50: $2.50 for input and $1.00 for output. This estimate uses the listed text-token rates; it does not account for image-token use or tool-specific charges. Actual cost depends on the workload and the rates applicable to the chosen processing option. (developers.openai.com)
Is GPT-4o a fit for a production application?
OpenAI documents GPT-4o as accepting text and image inputs and producing text output, including Structured Outputs. Its model page lists a 128,000-token context window and a 16,384-token maximum output. Those specifications describe documented limits; they do not establish that the model will meet your application’s quality, latency, or throughput requirements. The page marks GPT-4o deprecated, so confirm current availability before selecting it. (developers.openai.com)
Before budgeting for production, test the workload you intend to ship:
- Use representative prompts, input sizes, and image examples, if applicable.
- Review output quality against human-reviewed expectations and validate structured responses against your required schema.
- Measure token use, latency, and failure rates under realistic traffic.
- Confirm the API model ID, availability, and rates, and plan for a migration if GPT-4o is retired.
These are evaluation steps, not claims that GPT-4o has passed a production test. Your budget depends on request volume, average input and output tokens, caching, and any image or tool use.
Bottom line
ChatGPT is free to use, but GPT-4o is no longer available in ChatGPT. GPT-4o remains documented as an API model, with token-based charges, but it is deprecated. For free ChatGPT use, choose from the models currently offered in your account. For a production integration, verify the API model’s current availability and rates, then estimate and test using your actual workload.
Sources
Retiring GPT-4o and other ChatGPT models
Improving GPT-5.6 Sol in ChatGPT—and expanding access to GPT-5.6 Luna for free users