Claim Up to $500 in Free API Credits From ChatGPT Pro: Who Gets It, and What It Actually Buys

OpenAI announced on 28 October 2026 that eligible ChatGPT Pro and Business subscribers can now claim API credits worth up to $500 a month. The announcement appeared through OpenAI's developer community and its API credits documentation, and it was reported the same day by several independent outlets. The idea is simple: a paid ChatGPT subscription now also funds a monthly budget on OpenAI's developer platform, so a subscriber can build an app, an agent, or an internal assistant without paying a separate API bill for the first part of that work.

That distinction matters. ChatGPT credits and API credits are not the same thing. Your ChatGPT subscription pays for chat access in the app. API credits pay for model calls made through the developer platform, where you build your own products with an API key. OpenAI is now linking the two billing systems, but it is not merging them. You still need an API organization, you still need to claim the benefit, and you still need to understand what the credit does not cover before you build a product on it.

Who gets the credit, and how much?

The amount depends on your exact plan, not on how much you use ChatGPT. For individual Pro subscribers, the monthly credit equals the subscription price on all three Pro tiers. Pro 100 gets $100 a month, Pro 200 gets $200 a month, and Pro 500 gets $500 a month. Business works differently. Credits are pooled across the workspace: $20 per paid Standard seat, or $100 per paid Premium seat, up to a cap of $500 per workspace per monthly billing cycle. A workspace with five Standard seats and three Premium seats therefore generates $100 plus $300, which is $400. A workspace with ten Standard seats and four Premium seats would add up to $600 on paper, but the cap brings it back to $500. Workspaces that pay annually up front receive the whole year in one grant, capped at $6,000 for the annual cycle.

Eligibility has several conditions that are easy to miss. The subscription must have been active for at least seven days. The person claiming must own the ChatGPT subscription or the Business workspace, and must also own the API organization that receives the credits. The ChatGPT account and the API account must use the same email address. Free trials are excluded, and the published eligibility list does not include Plus, Free, Go, Enterprise, or Edu plans. If you are on Plus, this benefit does not apply to you, even if you also pay for API usage separately. Do not upgrade on the assumption that you will qualify until you have checked the current eligibility page for your plan and region, because OpenAI can change the list.

How do you claim it, step by step?

OpenAI's developer community announcement gives a short claim path. First, claim the API credits in ChatGPT billing, at the billing settings page in ChatGPT. Business users claim through workspace billing settings or the Admin Console, and confirm the API organization that will receive the credits. Second, use the OpenAI Developers plugin in ChatGPT or in Codex to create an API key and start a project. The billing screen shows the exact amount before you confirm, which matters because the first grant is prorated to the time left in the billing period you have already paid for. If you claim mid cycle, your first credit will be below the full monthly figure. That is normal, and it is not a bug or a missing payment.

After you claim, the credit is applied to API usage charges before any purchased credits you already hold. When the free credit runs out and no other balance remains, requests will fail with an exhausted credit balance error, rather than silently charging a card you have not added. In fact, one of the practical benefits is that subscribers can spend the grant without adding an API payment method at all. That makes it a genuine test budget for a first project, not just a discount on a bill you already pay. Keep a note of your billing cycle end date, because that date controls everything that follows.

What can you spend it on, and what is excluded?

The credit covers usage charges on the API platform for the Responses API, the Agents API, the Decisions API, the Realtime API including GPT-Live voice, and the Images API, plus built in tools such as web search, file search, and Code Interpreter, including their associated fees. That range is wide enough for most first projects: a document assistant that answers from your own files, a support triage agent that routes tickets, a voice prototype, an image feature, or a decision routing step built on the Decisions API.

The clear exclusion is Codex. Usage through an API key for Codex is not covered by this benefit, and the credit does not increase your usage limits for ChatGPT or for Codex. Do not budget a Codex workflow on this credit. The credit also does not roll over. Monthly grants expire at the end of the subscription billing cycle, and unused credit does not carry into the next month. Annual workspace grants expire at the end of the annual cycle. That expiry rule is the whole strategy question: a credit you do not use in the cycle is a credit you lose, so a small project that runs every week is worth more than a large project you plan to start next month.

What does $100, $200, or $500 actually buy in tokens?

Token prices vary by model, so a dollar figure alone tells you very little. Use published model rates to translate the credit into work, and treat every figure below as an illustration on stated rates, not as a promise of output quality or cost for your project. Two rates already published on this site give a useful range. The Decisions API, covered in an earlier TopUseAI post, costs $0.10 per million uncached input tokens, with no output charge. At that rate, $100 buys 1,000 million input tokens, which is 1 billion tokens, for decision routing work. That is why a ticket triage or classification agent is such a good first use for the credit: the per decision cost is tiny, and the credit stretches a long way.

At the other end, a general purpose model costs far more per token. GPT-6.1 Terra, covered in an earlier TopUseAI post, costs $1 per million input tokens and $5 per million output tokens, with cached input at $0.05 per million. At those rates, $100 buys 100 million input tokens if you spend it on input alone, or 20 million output tokens if you spend it on output alone. A realistic mixed job, for example 20 million input tokens and 8 million output tokens on Terra, costs $20 plus $40, which is $60, leaving $40 of a $100 credit for a second job or for retries. The lesson is not that one model is better. It is that the credit is generous for narrow, high volume work and modest for long, output heavy work, and you should pick the model that fits the job before you spend the grant.

What should you build first, and what should you avoid?

Build one small thing that runs every week, and that you can measure. A good first project is an internal assistant over your own documents, a support inbox triage step, or a lead qualification agent that reads an enquiry and drafts a structured summary. All three use the Responses or Agents APIs, all three benefit from file search, and all three produce a result you can check in minutes. Set a usage alert at 50 percent and at 80 percent of the credit, keep the first version to one model and one workflow, and log the token cost per finished task for the first twenty tasks. That log is what tells you whether the project is affordable after the credit, not the launch announcement.

Avoid three traps. First, do not build a customer facing product that depends on free credit to stay profitable, because the credit expires monthly, eligibility can change, and a customer contract needs a price you control. Price the product on paid API rates, and treat the credit as a test budget. Second, do not spend the whole grant on one long experiment in the last week of the cycle, because retries, long outputs, and tool calls can burn the balance before you have a result you can reuse. Third, do not assume the credit raises your ChatGPT limits. It does not. ChatGPT usage and API usage remain separate systems, and a Pro chat limit is not changed by claiming API credits.

If you are on Pro 100, $100 a month is a real budget for a side project, a prototype, or an internal tool, and it costs you nothing extra beyond the subscription you already pay. If you are on Business, the pooled cap rewards a small team that actually builds together, up to $500 a workspace, and penalises a large workspace that expected the credit to scale with every seat. Either way, claim it in billing, check the prorated first amount, build one measurable project, and use the credit before the cycle ends. A monthly benefit you forget to claim is the same as no benefit at all.

Key takeaways:

• Eligible ChatGPT Pro subscribers get $100, $200, or $500 in API credits per month, matching their Pro tier, and Business workspaces pool $20 per Standard seat or $100 per Premium seat up to $500 a workspace.
• The subscription must have been active for at least seven days, the claimant must own both the subscription and the API organization, and both accounts must use the same email.
• Claim in ChatGPT billing, or workspace billing for Business, and check the prorated first grant before confirming, because the first amount depends on time left in the cycle.
• Credits cover the Responses, Agents, Decisions, Realtime, and Images APIs plus built in tools, but they do not cover Codex usage and do not raise ChatGPT or Codex limits.
• Monthly credits expire at the end of the billing cycle and do not roll over, so build one small weekly project rather than planning a large project for next month.
• At $0.10 per million input tokens, $100 buys 1 billion Decisions API input tokens, while at Terra rates the same $100 buys 100 million input tokens or 20 million output tokens, so pick the model that fits the job.

Sources: OpenAI Developer Community announcement, Claim up to $500 in API credits using your ChatGPT Pro or Business subscription, 28 October 2026; OpenAI API credits documentation as reported by ChainCatcher and KuCoin, 28 to 29 October 2026; AlphaSignal, OpenAI Gives ChatGPT Pro Subscribers up to $500 in Free API Credits, 28 October 2026; CryptoBriefing, OpenAI offers up to $500 in monthly API credits to ChatGPT Pro and Business subscribers, 28 October 2026; earlier TopUseAI posts on the Decisions API and GPT-6.1 Terra pricing.

Affiliate disclosure: No affiliate link is used in this article. Links to OpenAI and to TopUseAI posts earn no commission.