ChatGPT Pro and Business now include up to $500 a month in API credits. Here's how to spend them on GPT-6.1 Terra

If you pay for ChatGPT Pro or Business, OpenAI is now giving you free API credits every month. On October 28 it announced that eligible subscribers can claim $100 to $500 a month on Pro, depending on the tier, and $20 or $100 per paid seat on Business, pooled up to $500 per workspace. The credits pay for real API usage: the same calls your n8n, Make or custom automations make with an OpenAI key.

Two days earlier, OpenAI also shipped GPT-6.1 Terra, a fast model priced at half of GPT-6.1 Sol. Together, they can cover a lot of small-team automation at low or no cost. But nothing arrives until you claim, and unused credits expire at the end of each billing cycle. Here's how to set it up so the money gets used.

What changed

API credits for subscribers. According to OpenAI's developer docs, the monthly amounts are:

Business credits are pooled and capped at $500 per workspace per month. Workspaces that pay annually upfront get the full year in one grant, capped at $6,000.

Plus, Go, Free and Enterprise are not eligible. Your subscription has to be paid, active and at least seven days old. Trials and complimentary plans don't count.

GPT-6.1 Terra. Released October 26. OpenAI describes it as "a fast, efficient model for visual reasoning, precise computer use, and clear, natural writing" in its API changelog. The API model ID is gpt-6.1-terra.

What Terra costs next to Sol

Prices per million tokens, from OpenAI's model pages for Terra and Sol:

Batch and Flex processing cost 50% less than standard on both models. Both also charge more for very long prompts: above 272K input tokens, input and cache rates are billed at 2x and output at 1.5x for the whole request.

Here is a rough example. Say an automation reads an inbound email and writes a short reply draft, using about 2,000 input tokens and 300 output tokens. On Terra that run costs about $0.0035. A $100 monthly grant covers roughly 28,500 of those runs. The same job on Sol costs about $0.007, so roughly 14,000 runs. This assumes no reasoning tokens. At the default medium effort, expect more output tokens and a higher cost, so test before you rely on these numbers.

Claim the credits first

To claim the credits, follow OpenAI's steps:

1. Sign in to ChatGPT with the account that owns the subscription.
2. On Pro, go to Settings, then Billing. On Business, go to the workspace billing settings or the Admin Console.
3. Find the API credits benefit and start the claim flow.
4. Pick the API organization that should receive the credits, then confirm.

You must own both the subscription and the API organization, and both accounts need the same email address. If you're new to the API, you can link your Personal API Organization.

Your first grant is prorated for the rest of your current billing cycle. After that, credits arrive automatically at each renewal.

Point your automations at the right organization

The credits only apply to the API organization you linked, and requests must use a project API key from that organization. If your n8n or Make workflows use an older key from a different organization, they will keep billing that organization's paid balance.

So after you claim:

• Create a new project API key in the linked organization.
• Swap it into each tool that calls OpenAI with your own key.
• Run one workflow and check the usage page to confirm the spend lands where you expect.

Pick the right model for each job

OpenAI's model guide splits the current lineup like this:

• GPT-6.1 Terra for visual reasoning, computer use and clear writing. OpenAI's example is matching product photos to catalog entries.
• GPT-6.1 Sol for complex work where you need close to top-tier performance at a lower price than Astra.
• GPT-6 Astra for the hardest reasoning, coding and professional work.
• GPT-6 Luna for focused, high-volume tasks where speed and cost matter most.

For most small-team automations (drafting replies, summarizing calls, tidying CRM notes, reading screenshots or invoices) Terra is a good first model to test. Keep Sol or Astra for the jobs OpenAI's guide reserves for them: complex, demanding reasoning and coding work.

Terra supports reasoning effort from none up to max, with medium as the default. Start at medium, then try low on simple tasks and compare the output. One technical note for builders: tool calling on Terra, Sol and Astra needs the Responses API, even though they also accept Chat Completions requests.

Know the limits

• Credits expire. Unused credits disappear at the end of each billing cycle. Annual Business grants expire at the end of the annual cycle.
• They don't cover Codex via API keys, and they don't raise your ChatGPT or Codex usage limits.
• When credits run out, requests fail unless you have purchased credits. You'll get an HTTP 429 with the error code credit_balance_exhausted. If a workflow matters, add paid credits or turn on Auto Recharge so it doesn't stop mid-month.
• Rate limits still apply. Your organization's normal API rate limits don't change.
• The claim isn't working for everyone yet. In OpenAI's community thread, several Pro users who pay through the App Store or Google Play say no claim option appears or the credits never arrived, and one Business user reported the same. Several say they have contacted support, but none report a fix yet.

What to do this week

1. Check your plan. If you're on Pro or Business and the subscription is at least seven days old, claim the credits now. The first grant is prorated, so the later you claim, the smaller it is.
2. Create a project API key in the linked organization and swap it into your automation tools.
3. Pick one workflow that runs often and switch it to gpt-6.1-terra. Run 20 real examples and compare the output with your current model.
4. Set a reminder a few days before your renewal date to check how much credit is left. If you're leaving most of it unused, move another workflow over.
5. If a workflow can't afford to fail, add a small paid balance or Auto Recharge as a backstop.