Your Monthly Plan Now Ships With an "API Wallet"── Claude's Max and Team Plans Let You Run Your Own Agents on a Subscription (Interactive Claude Code Not Included)
Claude's Max and Team plans now include monthly API credits. Max 20x gets $200, Team up to $500. The credits go toward running your own agents via the Agent SDK or claude -p, but the limits are clear: interactive Claude Code and rollover are both out of scope.
Subscriptions Have Started Handing Out a Monthly "API Balance"
On October 7, 2026, Anthropic announced that Claude's Max and Team plans will now include monthly API credits. Until now, the monthly subscription covered the "human-at-the-screen" use cases — chatting on claude.ai and interactive sessions in Claude Code. With this change, the same subscription now adds a balance for calling the Claude API programmatically. You can start running your own apps and agents without registering an additional payment method.
Put simply, the monthly plan now comes with an "API wallet" built in from the start. That's one less hoop for developers who used to set up separate usage-based API billing just for testing or small-scale automation.
Monthly Credits by Plan
The amount granted is determined by your plan. For Team plans, the amounts for Standard and Premium seats are summed and pooled into a single balance, capped at the equivalent of $500 per month.
| Plan | Monthly credits |
|---|---|
| Max 5x | $100 |
| Max 20x | $200 |
| Team (Standard seat) | $20 per seat |
| Team (Premium seat) | $100 per seat |
| Discounted Team (nonprofit / researcher) | Same as Team (Standard $20 / Premium $100) |
Seat counts are calculated at the start of each billing month. For example, a team with 3 Standard seats plus 2 Premium seats gets $260 a month; add one more Premium seat and the following month becomes $360 (both within the $500 pool cap). The Free, Pro, and Enterprise plans are not eligible.
The Credits Are for "Programmatic" Use── Interactive Claude Code Is Not Included
This is the most important caveat. These credits apply only to API usage on the Claude platform (Console). It's easy to get confused, but interactive Claude Code sessions in your terminal are not covered. If you launch Claude Code from an organization that holds only credits, you'll see an insufficient-balance message.
| Use case | Eligible for credits |
|---|---|
| Claude API (Messages / Batches) | ◯ |
| Claude Agent SDK | ◯ |
| claude -p (run with the linked organization's API key) | ◯ |
| Claude Managed Agents | ◯ |
| Console Playground | ◯ |
| Interactive Claude Code sessions | ✗ |
| Overages on the Claude app | ✗ |
| Other platforms such as Bedrock / Vertex / Microsoft Foundry | ✗ |
The flip side is that unattended, batch-style automation using the Claude Agent SDK or claude -p is within the credits' reach. A setup that hits the API from a script to generate articles or crawl sources is exactly the kind of usage these credits are meant to cover.
How to Receive Them: Link Exactly One Console Organization
They aren't applied automatically. From "Settings > Billing (Max)" or "Organization settings > Billing (Team)" on claude.ai, link the Claude Console organization that will receive the credits. The granted amount appears under "Settings > Billing" in the Console, in the "Promotional credits" section, along with its value and expiration date. Receiving credits requires 7 days since you joined the plan, plus Owner, Admin, or Billing permissions on the Console side.
Worth noting: the link is one-to-one, and you can't reassign it yourself later. One plan can only be linked to one Console organization, and each Console organization can only receive from one plan. Changing it requires contacting support, so it's safest to choose "the organization you'll actually develop in" from the start.
What Changes for Developers in Practice
The biggest effect is that the upfront cost of "just trying things" goes down. Subscribers can run an agent built with the Agent SDK or a small automation and get a feel for it, without separately signing up for usage-based API billing. With Max 20x that's a $200 monthly experimentation budget, and with Team up to $500 — no extra payment required.
There's a boost on the rate-limit side, too. An organization with linked credits is bumped to at least the Start tier, and to the Build tier if its monthly credits exceed the equivalent of $200 (though the credits themselves don't count toward the qualifying conditions for a higher tier). It's designed to make it easier to move a prototype one step forward.
Pitfalls That Are Easy to Miss
- Expires monthly, no rollover: Unused credits disappear at the end of the billing cycle (monthly for annual plans). Assume you can't stockpile them, and plan to use them up.
- Capped at $500: No matter how many seats a Team adds, the pool tops out at $500. Steady production use will require separately purchased credits.
- Runs out and stops: Once both the granted credits (consumed first) and purchased credits are exhausted, the API halts until next month's credits arrive. On the plus side, your plan itself isn't charged, so there's no "surprise large bill" — but there's a risk that your automation quietly stops. Auto-reload applies only to purchased balances.
- Rigid linking: It's fixed at one plan = one organization, and reassigning requires a support request.
On the whole, this change widens the door for subscribers to casually try out their own agents. At the same time, the boundaries — "no rollover, there's a cap, and interactive Claude Code is separate" — are clear. If you budget on the assumption that these credits can fully cover the cost of steady production operation, you could get tripped up by end-of-month expiration or a run-dry shutdown. The closer reading is to treat them as a budget for testing and small-scale operation.
References: API credits for Max and Team plans (Claude Platform Docs) / Monthly API credits for Max and Team plans (Claude Help Center) / Claude Platform release notes (2026-10-07)
