claude max vs team api credits: how much you get, and what they will and will not pay for
on october 7, 2026, anthropic started adding monthly api credits to claude max and team plans. if you already pay for one of those plans, you have a budget for your own apps and agents that you may not have claimed yet. the fine print decides whether it is useful.
claude max and team plans now include monthly claude api credits: $100 on max 5x, $200 on max 20x, and $20 or $100 per team seat, pooled up to $500. they cover the api, managed agents, and the agent sdk. they do not cover interactive claude code, do not roll over, and simply stop when spent.
that is the whole map. the rest of this note is what each piece means for a small business or a solo builder: who gets what, the two exclusions people will trip on, what happens at the end of the month, and a five-step setup that points the credits at one job worth automating.
what anthropic actually changed
max and team subscribers can now claim a monthly api balance on the claude platform, separate from their chat and claude code limits. anthropic's release notes for october 7, 2026 describe it as:
"monthly API credits for running your own apps and agents on the Claude Platform"
anthropic, claude release notes, october 7, 2026
the credits roll out over a few days and you claim them by linking a claude console organization from your plan's billing settings (anthropic release notes, oct 7 2026). until now, a claude subscription and the claude api were two separate bills. a pro or max plan bought you chat, cowork, and claude code; anything you built yourself on the api was paid for in the console. this change does not merge the two, but it puts a free monthly api allowance on the higher plans.
it matters more this month because the other route founders were reaching for is shut. anthropic's claude startups page says its current cohort is full and it is not accepting new applications, and points people to a newsletter for the next round (claude startups page, accessed oct 15 2026). the plan credits are not gated by an application. if you are on max or team, you qualify after seven days.
how much each plan gets
max 5x gets $100 a month, max 20x gets $200, and team gets $20 per standard seat plus $100 per premium seat, pooled and capped at $500. every figure below is from anthropic's help article on the credits (anthropic help center, accessed oct 15 2026).
| what you are comparing | max 5x | max 20x | team |
|---|---|---|---|
| monthly api credit | $100 (anthropic help center, oct 2026) | $200 (anthropic help center, oct 2026) | $20 per standard seat, $100 per premium seat, pooled, capped at $500 (anthropic help center, oct 2026) |
| who claims it | the subscriber | the subscriber | one owner or primary owner, for the whole team |
| wait before claiming | seven days on the plan | seven days on the plan | seven days on the plan |
| where it lands | one console organization, which you cannot change yourself | one console organization, which you cannot change yourself | one console organization, shared by every api key in it |
| rollover | none, expires each billing cycle | none, expires each billing cycle | none, expires each billing cycle |
| best fit | one solo builder, one or two recurring automations | a heavier solo builder or several agents | a small team sharing one budget, with spend limits per project |
anthropic's own team example is three standard seats and two premium seats receiving $260 a month (anthropic help center, accessed oct 15 2026). the pool is recalculated from the seats on the plan each time credits are deposited, so adding a premium seat raises the next month's balance until you hit the $500 cap. free, pro, and enterprise plans are not eligible (anthropic help center, accessed oct 15 2026).
what the credits pay for, and the two things they do not
the credits pay for building your own things on the claude platform; they do not pay for using claude code interactively, or for extra usage once you hit your plan limits. covered: the messages api and message batches api, the console playground, claude managed agents, and the claude agent sdk, on any available claude model (anthropic help center, accessed oct 15 2026).
the two exclusions are where people will get caught:
- interactive claude code is not covered. terminal, ide, desktop, and web sessions keep drawing from your plan's usage limits, not from the credits (anthropic help center, accessed oct 15 2026). the credits do not buy you more claude code.
- extra usage is not covered. if you hit your chat or claude code limits and turn on extra usage, that is billed separately; the api balance does not top it up (anthropic help center, accessed oct 15 2026).
there is one useful edge case for builders. claude -p and the agent sdk are covered when you run them yourself with an api key from the linked console organization, because that is billed as agent sdk usage. signed in with your claude plan instead, they still count against your plan limits. runs started by the claude code github action, an ide extension, or the desktop app count as claude code usage and are not covered, even with -p (anthropic help center, accessed oct 15 2026). if you have a headless script on a schedule, give it an api key from the linked organization and it runs on the credits.
what happens when the credits run out
nothing is charged to your claude plan; if the console organization has no other credits, api requests simply stop until next month. if the organization has purchased credits or auto-reload turned on, usage continues and draws from those, and the monthly credits are always spent first (anthropic help center, accessed oct 15 2026).
for a business, "stops" is the important word. an automation that classifies inbound leads or drafts support replies will go quiet on the day the balance hits zero, and nobody gets an error in their inbox unless you built one. so decide up front, per job: is it fine for this to pause until next month, or does it need auto-reload behind it? on team plans, every api key in the linked organization spends the same pool, so set workspace spend limits in the console before a teammate's experiment eats the month (anthropic help center, accessed oct 15 2026).
how far the credits actually go
far, if you send short, repetitive jobs to a small model, and not very far if you send everything to the biggest one. the credits work with any available claude model, so the model you choose sets how many jobs a month fit inside them. claude haiku 5.5 costs $0.10 per million input tokens and $0.50 per million output tokens on prompts up to 100,000 tokens (anthropic, oct 7 2026), a twentieth of sonnet 5.5's per-token price. i broke down which small-business jobs belong on which model in claude haiku 5.5 vs sonnet 5.5, and the general rule is in you are using one ai model for everything: route by job shape, not by habit.
the honest constraint is not the dollar figure. it is that most small businesses do not yet have a single recurring job built on the api. the credits are free capacity with nothing pointed at it. if you are deciding what that first job should be, ai agents for small business covers where they earn their keep, and claude code mods vs hooks vs skills covers what to build inside claude code instead, where these credits do not reach.
the five-step setup
claiming takes a few minutes; picking the job is the part that makes the credits worth anything.
- check you are eligible. active max or team plan, held for at least seven days. on max you must be the subscriber; on team, the owner or primary owner (anthropic help center, accessed oct 15 2026).
- pick the console organization carefully. you can link only one, and you cannot change it yourself afterwards. use the one you will actually build in, signed in with the same email as claude.ai.
- claim. on claude.ai, settings > billing (max) or organization settings > billing (team), then "link organization" in the api credits section, review the supplemental credit terms, and confirm.
- point it at one recurring job. one bounded task that runs every week: lead triage, ticket summaries, a weekly report. build it on the api, managed agents, or the agent sdk with a key from the linked organization.
- decide the run-out rule. pause until next month, or auto-reload. write it down next to the job, so a quiet automation in week four is a choice and not a surprise.
the takeaway
if you are on max or team, you have a monthly api budget whether you use it or not, and it expires every cycle. it will not make claude code go further, and it will not bill you when it runs out; it just stops. the value is not the credit itself. it is finally having a reason to build the one small automation you kept postponing because the api was a separate bill. claim it, point it at one job, and decide what happens on the day it hits zero.
build the first job the credits pay for
the ai builder toolkit is a set of claude skills that encode how the work gets done in a one-person business: which jobs to hand to claude, how to scope them so a small model runs them reliably, and how to check the output before it ships.
see the toolkitnewer to this and want the version with people in the room? the ai builders lounge is where the weekly builds happen.
or just follow along. new field notes most weeks on x, instagram, and tiktok.