
AI-generated editorial illustration; not a screenshot.
What the Help Center Says
This article reflects Anthropic's Help Center article on monthly API credits as I reviewed it on October 8, 2026, and the October 7 release notes that announced it. Terms can change, so check the live page before you plan around them.
As of that date:
- Max 5x includes $100 per month in Claude Platform API credits.
- Max 20x includes $200 per month.
- Team plans receive pooled credits based on their seats.
- Free, Pro and Enterprise plans are not eligible for this credit.
These are separate API credits for your own apps and agents. They are not extra chat capacity.
The Conditions That Decide Whether You Get Them
A headline amount is not the same as a balance in your account. Several conditions apply:
- Eligibility and plan status. The plan must be active and in good standing.
- Claiming. You may need to claim the credits in Claude Settings, under Billing, and link a Claude Console organization. The credits go to that organization.
- New subscribers. A new subscriber needs seven days on an eligible plan before claiming.
- Rollout. Anthropic describes a rollout over a few days, so availability on any one account is not guaranteed on any one day.
- Expiry. Credits refresh each billing cycle, and unused credits expire at the end of that cycle. They do not roll over.
I do not know whether my own account is eligible for a balance right now, whether I have claimed one, or how much is unused. I am not claiming any of that, and you should not assume it for yours.
What the Credits Cover, and What They Do Not
According to the help article, the credits cover the Claude API, the Console Playground, Claude Managed Agents and the Claude Agent SDK. They do not cover interactive Claude Code, extra usage in Claude, Claude Code or Claude Cowork, or Claude accessed through Bedrock, Vertex AI or Microsoft Foundry.
The most important practical point: these credits do not increase your interactive Claude or Claude Code limits. If the reason you want more is that you keep hitting a chat or coding limit, this does not change that. If you want to run your own scripts, evaluations or agents against the API, it may.
Making Use of an Allowance That Expires
Use-it-or-lose-it allowances reward planning. A short routine works:
- Find your reset date in the billing page and put it in a calendar.
- Decide on one useful experiment before the cycle starts: an evaluation run over your own test cases, a batch classification job, a prototype of an internal agent.
- Set a cap so an experiment cannot become a surprise charge once credits are gone. Confirm in Console how spend beyond credits is handled for your organization.
- Record the result, including cost, time and whether the output was acceptable. A credit spent on something nobody reviews is still wasted.
For a team, assign an owner. A pooled allowance with no owner is the one most likely to expire.
Why I Track This
I spend around $600 a month across AI subscriptions, which is my own estimate rather than an audited bill. I built a private dashboard, described in the usage meter article, to see accounts, models and usage in one place. In the original post I described tracking API credits as something I wanted to keep visible; that was a consideration, not a shipped feature.
For another personal data point about choosing between tools, see my Claude and Codex comparison. It rests on informal tests and should be read that way. If you are planning an API experiment with spare credits, the Haiku 5.5 subagent plan is one candidate.
Read Next
The original post, later edited to sharpen the expiry point, is on LinkedIn. The primary source is the Claude Help Center article. The Agent companion, How to put an expiring API allowance to work, looks at the operating side: budgets, caps and review.
If you want an experiment designed with acceptance criteria rather than improvised, our AI agent development service can scope it. Questions are welcome through the contact page.