Short answer: There is no current ChatGPT subscription price for the original GPT-4 model: GPT-4 was retired from ChatGPT effective April 30, 2025. OpenAI lists ChatGPT Plus at $20 per month, but that is the subscription price for ChatGPT—not a way to restore access to original GPT-4 or pay for API calls. OpenAI bills ChatGPT and API use separately. (community.openai.com)
For API use, OpenAI’s GPT-4 model page marks the model shut down. That means its displayed catalog rates are not a usable price quote for new production calls. GPT-4, GPT-4o, and GPT-4o Mini are distinct model names; don’t transfer one model’s price to another.
ChatGPT subscription vs. GPT-4 API cost
| What you mean | Price or status | What it means for your budget |
|---|---|---|
| Using ChatGPT | ChatGPT Plus is listed at $20/month. | This is a monthly ChatGPT subscription price, not a GPT-4-specific fee or API allowance. Original GPT-4 is no longer available in ChatGPT. Check OpenAI’s current plan information for what models your account can use. (OpenAI: ChatGPT Plus) |
| Calling original GPT-4 through the API | OpenAI’s GPT-4 model page marks it shut down. | There is no usable current rate for new GPT-4 production calls. Do not treat a price displayed on a shut-down model page as a current quote. (OpenAI: GPT-4 API model) |
| Using the API generally | API billing is separate from ChatGPT subscriptions. | Estimate API charges from the exact model, its applicable rates, and your input and output usage. (OpenAI: ChatGPT and API billing) |
Bottom line: Paying $20/month for Plus does not buy original GPT-4 access or standard API usage. The amount is the ChatGPT Plus subscription price; the model access and API bill are separate questions.
Budgeting for a production API alternative
If you need a current API model for work that used GPT-4.x, OpenAI’s migration guidance recommends GPT-6 Luna as a starting point and advises evaluating representative requests before moving production traffic. That recommendation is a starting point for evaluation, not a guarantee that Luna will match your application’s results. (OpenAI: GPT-4.x migration guidance)
The API model identifier is gpt-6-luna. It accepts text and image input and returns text; its context window is 1,050,000 tokens, with a maximum output of 128,000 tokens. Audio and video are not supported by this model. For a text-only workload that fits those limits, its documented input/output types and limits are a concrete starting point for assessing fit. (developers.openai.com)
OpenAI lists GPT-6 Luna’s standard price as $0.10 per million input tokens and $0.50 per million output tokens, for requests with input lengths under 272,000 tokens. Batch and Flex processing are priced at half the Standard rates; Fast mode costs twice the applicable rates. (developers.openai.com)
For an illustrative month with 10 million input tokens and 2 million output tokens, assuming Standard pricing and each request stays below the 272,000-input-token threshold:
- Input: 10 × $0.10 = $1.00
- Output: 2 × $0.50 = $1.00
- Estimated model-token total: $2.00
This is a calculation from the listed Standard rates, not a measured test or a guarantee that Luna will perform adequately for your use case. It excludes tools and other charges. If a request exceeds the stated input threshold, the model page lists higher rates for that request. (developers.openai.com)
Before you move production traffic
- Confirm the purchase type: ChatGPT interface or API. They are billed separately.
- Check the exact model identifier, current availability, and rate mode in OpenAI’s API catalog; don’t budget with a price shown for a shut-down model.
- Check workload fit: confirm your input type, context needs, and expected output length fit the model’s documented capabilities and limits.
- Test representative prompts and edge cases. Compare output quality, latency, cost, formatting, and any required tool use.
- Roll out gradually and monitor results. OpenAI’s migration guidance recommends evaluating your workload before completing a migration. (OpenAI: migration guidance)
FAQ
Is ChatGPT Plus the GPT-4 price?
No. Plus is listed at $20 per month as a ChatGPT subscription. Original GPT-4 was retired from ChatGPT in 2025, so the Plus price is not a fee for access to that model. API usage is billed separately. (community.openai.com)
What is the current GPT-4 API price?
There is no usable current price for new GPT-4 API calls: OpenAI’s model page marks GPT-4 shut down. Any rates displayed for the shut-down model should not be treated as a new production quote. (OpenAI: GPT-4 API model)
Can I use a GPT-4 rate to estimate GPT-4o costs?
No. GPT-4 and GPT-4o are different model names. Check the exact model identifier and its own current API pricing rather than transferring rates between models.
Sources
ChatGPT Release Notes — OpenAI Help Center
Sunsetting GPT-4 in ChatGPT (not in API) — OpenAI Developer Community
What is ChatGPT Plus? — OpenAI Help Center
Managing billing for ChatGPT and the API platform — OpenAI Help Center