API Credits Now Included in ChatGPT and Claude Plans: How to Claim and Use Them

Last updated: 9 November 2026

OpenAI: API credit with ChatGPT plans

Rules to know:

• The API organisation must belong to the subscription or workspace owner, using the same email address. A Business admin role is not enough.
• Each subscription funds one API organisation, and each organisation takes credit from one subscription.
• Free credit is used before any purchased credit. When both run out, requests fail unless you have a payment method.
• Codex usage billed through API keys is excluded.
• From 30 October 2026, OpenAI halved the Codex and ChatGPT Work usage included in Pro 200, and the API credit does not offset that.

Anthropic: API credit with Claude plans

Rules to know:

• Interactive Claude Code in the terminal, IDE, desktop or web is excluded, as is extra usage in Claude, Claude Code or Cowork.
• Claude through Amazon Bedrock, Google Cloud Vertex AI or Microsoft Foundry is excluded.
• Agent SDK runs or claude -p using an API key from the linked Console organisation can use the credit.
• Claim after seven days on an eligible plan: Settings, then Billing for Max and Ultra; Organization settings, then Billing for Team owners.
• Link one Claude Console organisation. The Team credit is claimed once for the whole team.

Anthropic also renamed Max 20x to Ultra in October 2026. Verified Ultra accounts no longer have a five-hour usage limit, according to reporting on the change. See Claude Pro vs Max and Claude usage limits.

How to claim your credit

1. Check eligibility. Make sure you are on a qualifying plan and have been for at least seven days.
2. Create or find your API organisation. OpenAI Platform or Claude Console, using the same email as the subscription owner.
3. Open billing settings in ChatGPT or Claude and claim the credit.
4. Choose the organisation that should receive it.
5. Create an API key in that organisation.
6. Set a usage alert so you know when the credit is running out.

What you can do with $100 to $500 a month

For a small business, this is real capacity. Some practical uses:

Make the credit go further

• Pick the cheapest model that does the job. Many tasks do not need a frontier model. See the cheapest AI model for each task.
• Use prompt caching for repeated instructions. See prompt caching explained.
• Use batch processing for non-urgent work at about half price. See batch API savings.
• Keep prompts short and outputs structured. See how to reduce token usage.
• Estimate before you build. See how to estimate AI API costs.

Does this change which plan you should buy?

The credit makes higher plans better value if you would use the API anyway. A few scenarios:

• A solo professional on Claude Pro who rarely uses the API gains nothing by upgrading just for credit.
• A developer already paying for API usage and a Max or Pro plan effectively gets part of the subscription back.
• A small team on ChatGPT Business with five Standard seats gets $100 a month of API credit at annual billing, enough for several automations.
• Anyone relying on Claude Code should note the credit does not cover interactive Claude Code use. See how to reduce Claude Code costs.

See AI subscriptions worth paying for and how to save money on AI subscriptions.

Common mistakes

• Linking the wrong organisation or one owned by a different email address.
• Expecting the credit to raise ChatGPT or Claude app limits. It does not.
• Letting it expire unused. Set a reminder to put it to work each month.
• Running out with no payment method and having production automations fail. Decide whether to add a card with a spending limit.
• Sharing API keys widely. Keep keys secret and rotate them.

Example: turning $100 of credit into a useful automation

An illustrative example for a small online store on ChatGPT Business with five Standard seats ($100 of monthly credit on annual billing):

1. Goal: draft product descriptions for new arrivals and tag incoming customer emails.
2. Setup: an n8n workflow watches a spreadsheet of new products and a support inbox.
3. Descriptions: for each new product, the workflow sends specifications to a mid-priced model and saves a draft description for review.
4. Email tagging: each new email is sent to a cheap model that returns a category and urgency, which the helpdesk uses for routing.
5. Cost control: a usage alert at 80% of the monthly credit, and a rule to pause non-urgent tasks if credit runs low.

Hundreds of short classification requests and dozens of descriptions a month typically fit comfortably within $100 on cost-efficient models. Measure your own usage in the first month.

Security basics for API keys

• Create separate keys for separate projects so you can revoke one without breaking others.
• Never paste keys into public code, shared documents or chat.
• Store keys in a secrets manager or your automation tool’s credential store.
• Set project-level spending limits where the platform allows it.
• Rotate keys if a team member leaves or a key may have been exposed.

Comparing the two offers

See OpenAI API pricing and Claude API pricing for model rates.

Frequently asked questions