OpenAI gives ChatGPT Pro and Business subscribers up to $500 a month in API credits

OpenAI has turned paid ChatGPT subscriptions into a source of monthly developer budget. Since October 28, 2026, eligible ChatGPT Pro and ChatGPT Business subscribers can claim free credits on the OpenAI API Platform every billing cycle, worth up to $500 per month. The change was recorded as a feature entry in the official OpenAI API changelog on that date and is documented in a dedicated developer guide, "API credits for ChatGPT subscribers", which sets out the amounts, the eligibility rules and the way the credits are consumed.
The amounts depend on the plan. Pro subscribers receive $100, $200 or $500 a month depending on their Pro tier, while Business workspaces receive $20 for every paid Standard seat and $100 for every paid Premium seat, pooled at workspace level and capped at $500 a month. The credits are not a discount on ChatGPT itself and they do not raise any ChatGPT or Codex usage limit: they are a prepaid balance on a separate API organization that the subscriber owns, intended, in OpenAI's own words, for building apps and agents.
For individual developers and small teams who already pay for ChatGPT, the practical effect is that a meaningful amount of API usage no longer requires a separate monthly top-up. For OpenAI, the benefit ties its consumer and workplace subscriptions more closely to its developer platform, a link that Anthropic established three weeks earlier when it began bundling monthly API credits with Claude Max and Team plans.
··········
HOW MUCH EACH CHATGPT PLAN RECEIVES.
The credit schedule published by OpenAI is short, but the Business side has a pooling mechanism that determines how much a workspace actually receives. Seat credits are added together across the workspace and then capped, which means larger teams reach the ceiling quickly.
........
ChatGPT plan | Monthly API credit | How it is calculated |
Pro | $100, $200 or $500 | Depends on the Pro tier |
Business, Standard seat | $20 per paid seat | Pooled per workspace |
Business, Premium seat | $100 per paid seat | Pooled per workspace |
Business workspace cap | $500 | Maximum per workspace per month |
Enterprise | Not eligible | Moving to Enterprise removes the link |
Plus, Go, Free | Not eligible | Downgrading removes the link at cycle end |
........
The arithmetic for Business workspaces follows directly from these figures. A workspace with 25 paid Standard seats already reaches the $500 ceiling, and so does a workspace with five paid Premium seats, so for any organization above that size the API credit stops growing with headcount. A mixed workspace with three Standard seats and two Premium seats would receive $260 a month. Business customers that pay annually up front receive the full year's grant in a single issuance rather than month by month, according to the guide.
The first grant is prorated for the remainder of the current billing cycle, so a subscriber who claims the benefit halfway through a cycle receives roughly half of the monthly amount at first. Upgrading a Pro tier or adding Business seats can add prorated credits within the caps, while seat reductions take effect from the next cycle and reassigning an existing seat to another person does not create new credits.
··········
WHO IS ELIGIBLE AND HOW TO CLAIM THE CREDITS.
OpenAI's rules are designed to stop the benefit from being used as a free API allowance on throwaway accounts. The ChatGPT subscription must be active, paid and at least seven days old, and free trials and complimentary subscriptions do not qualify. The person claiming must own both the ChatGPT subscription or Business workspace and the API organization that receives the credits, and the ChatGPT account and the API account must use the same email address.
The claim itself happens inside ChatGPT. Pro subscribers open their billing settings, while Business owners use the workspace billing settings or the Admin Console; in both cases they find the API credits benefit, select the API organization that should receive the money and confirm. Users who have never set up an API account can link the subscription to a Personal API Organization created for the purpose. Once claimed, the balance appears in the billing settings of the API Platform. In its announcement on the OpenAI Developers account, the company also pointed to the OpenAI Developers plugin in ChatGPT or Codex as a way to create an API key for the linked organization.
The link is deliberately rigid. Grants apply only to the linked organization, and requests must use one of that organization's project API keys. To change which organization receives the credits, a subscriber has to contact OpenAI support. Switching from Pro to Business ends the Pro link, and a Business owner can create a new link only after seven days.
··········
WHAT THE CREDITS COVER AND WHAT THEY DO NOT.
The coverage is broad on the API side. According to the guide, the credits pay for paid API Platform usage billed to the linked organization, including the Responses API, the Agents API, the Decisions API, the Realtime and GPT-Live APIs and the Images API, together with built-in tools and the storage charges those tools generate. In practice, this means the credits can fund the current GPT-6 family on the API, from the low-cost GPT-6 Luna used in the Decisions API to GPT-6.1 Terra, which OpenAI released on October 26.
The main exclusion concerns coding agents. Codex usage through API keys is not covered by the benefit, and the credits do not increase the usage limits that come with ChatGPT or Codex subscriptions. A developer who expected to stretch Codex further with the new balance will find that the two budgets stay separate.
The credits are also strictly monthly. Unused amounts expire at the end of each billing cycle and do not roll over, so a developer cannot accumulate several months of credit for a larger project later. When both free and purchased credits are present, the free credits are consumed first.
··········
WHAT HAPPENS WHEN THE MONTHLY CREDIT RUNS OUT.
Developers building on the benefit need to plan for exhaustion, because the behaviour depends on what else is in the account. If the organization still holds purchased credits, API usage simply continues and draws on that paid balance. If nothing remains, requests fail with an HTTP 429 response carrying the error code credit_balance_exhausted, and service resumes only after the owner adds paid credits or enables Auto Recharge.
That 429 code matters for application design. A 429 is the same status family used for rate limiting, so clients that blindly retry on any 429 may keep hammering the API after the monthly grant is gone. Applications that rely solely on the free credit should check the error code and stop or degrade gracefully rather than treating the failure as a temporary throttle. The benefit is also subject to OpenAI's API terms, usage policies and the organization's normal rate limits, so the credits do not raise throughput by themselves.
Cancelling a subscription does not claw back credits that have already been issued, but new grants stop as soon as the subscription becomes ineligible.
··········
WHAT $100 OR $500 BUYS AT CURRENT OPENAI API PRICES.
To translate the credit into workload, Data Studios applied the standard prices listed on OpenAI's official API pricing page on October 30, 2026 to two typical scenarios. The first is a customer-support chatbot request with 3,000 input tokens, of which 60% hit the prompt cache, and 500 output tokens. The second is a coding-agent session with 400,000 input tokens, 80% of them cached, and 40,000 output tokens. These assumptions are illustrative; real workloads vary widely with prompt length, cache behaviour and output size.
........
Model | Input / cached / output per 1M tokens | Chatbot requests per $100 | Chatbot requests per $500 | Agent sessions per $500 |
GPT-6 Luna | $0.10 / $0.01 / $0.50 | about 257,700 | about 1,288,600 | about 16,000 |
GPT-6.1 Terra | $1.00 / $0.05 / $5.00 | about 26,400 | about 131,900 | about 1,690 |
GPT-6.1 Sol | $2.00 / $0.10 / $10.00 | about 13,200 | about 66,000 | about 845 |
GPT-6 Astra | $10.00 / $1.00 / $50.00 | about 2,580 | about 12,900 | about 160 |
........
The calculation for GPT-6.1 Terra shows how the numbers are built. Each chatbot request costs 1,200 uncached input tokens at $1.00 per million ($0.0012), 1,800 cached tokens at $0.05 per million ($0.00009) and 500 output tokens at $5.00 per million ($0.0025), for a total of about $0.0038. A $100 grant therefore covers roughly 26,400 such requests a month, or about 880 a day across a 30-day month. The coding-agent session costs about $0.30 on Terra and about $3.12 on GPT-6 Astra, which means the top $500 grant funds roughly 1,690 Terra sessions but only about 160 Astra sessions.
For a solo developer prototyping an agent or running a side project, $100 to $200 a month on Terra or Luna is enough for steady testing and a modest production load. For a Business workspace, the $500 cap is meaningful for internal tools and pilots but small relative to a production application with thousands of daily users, where the credits act as a subsidy rather than a full budget.
··········
HOW THE OFFER COMPARES WITH ANTHROPIC'S CLAUDE API CREDITS.
The structure of OpenAI's benefit closely mirrors the monthly API credits that Anthropic introduced for Claude Max and Team plans on October 7, 2026. According to Anthropic's documentation, Max 5x subscribers receive $100 a month and Max 20x subscribers $200, Team Standard seats carry $20 each and Team Premium seats $100 each, and Team seats pool into one balance capped at $500 a month. The seven-day waiting period for new subscribers, the expiry at the end of each cycle without rollover, and the rule that included credits are spent before purchased ones are the same on both sides.
........
Item | OpenAI (ChatGPT) | Anthropic (Claude) |
Individual plans | Pro: $100, $200 or $500 | Max 5x: $100; Max 20x: $200 |
Standard team seat | $20 per Business Standard seat | $20 per Team Standard seat |
Premium team seat | $100 per Business Premium seat | $100 per Team Premium seat |
Team cap | $500 per workspace | $500 per organization |
Rollover | No | No |
Coding agent excluded | Codex via API keys | Claude Code |
Enterprise plans | Not eligible | Not eligible |
........
The clearest difference is at the top of the individual range. OpenAI's highest Pro tier carries a $500 monthly grant, while Anthropic's largest individual grant is $200 on Max 20x. Anthropic, for its part, documents an extra effect on rate limits: organizations linked to an eligible Claude plan move up to at least the Start usage tier, or at least Build when the monthly credit exceeds $200. OpenAI's guide says only that the benefit is subject to the organization's rate limits. Both companies exclude their own coding agents from the credits, which keeps Codex and Claude Code inside their subscription quotas rather than turning the API grant into extra agent capacity.
··········
WHAT IS ESTABLISHED AND WHAT REMAINS UNCLEAR.
The core facts are documented by OpenAI: the benefit went live on October 28, 2026; Pro subscribers receive $100, $200 or $500 a month depending on tier; Business workspaces receive $20 per Standard seat and $100 per Premium seat up to $500; credits expire monthly, are spent before purchased credits and cover the main API Platform products but not Codex through API keys. The eligibility conditions, the single-organization link and the credit_balance_exhausted error behaviour are written into the developer guide.
Several practical points are less clear. OpenAI's public pricing page does not currently show the dollar price of each Pro tier next to its credit amount, so subscribers have to check their own billing settings to see which grant applies. The guide does not spell out whether subscriptions purchased through mobile app stores qualify, a point Anthropic addresses explicitly for Claude Max. And OpenAI has not said whether the benefit is permanent or whether the amounts may change. For now, the offer gives paying ChatGPT users a recurring developer budget that, at current GPT-6.1 Terra and GPT-6 Luna prices, is large enough to run real prototypes and small production workloads without a separate API bill.
··········
FOLLOW US FOR MORE.
·····
DATA STUDIOS
·····
datastudios.org




