Anthropic adds monthly API credits to Claude Max and Team plans
Anthropic said on October 7 that Max and Team plans now include monthly Claude Platform API credits: $100 (Max 5x), $200 (Max 20x), up to $500 pooled for Team.
On October 7, 2026, the same day it released Claude Haiku 5.5, Anthropic announced that Claude Max and Team subscriptions will include monthly credits for the Claude Platform, its API developer product. Max 5x gets $100 a month, Max 20x gets $200, and Team gets up to $500, pooled across the team. The money can only be spent on apps and agents you build yourself: it doesn’t cover Claude Code, and unused credits expire each billing cycle.
Key points
- Max 5x and Max 20x credits ($100 and $200) match each plan’s price. Team gets $20 per Standard seat and $100 per Premium seat, pooled into one balance capped at $500.
- The credits cover the Claude API, Claude Managed Agents, the Claude Agent SDK and the Console Playground. They don’t cover interactive Claude Code, extra usage inside Claude’s apps, or Claude on Bedrock, Vertex AI and Microsoft Foundry.
- Credits arrive once per billing cycle and don’t roll over. When they run out, API requests stop; nothing is charged to the subscription.
Background
This offer follows a stretch of back-and-forth over how subscriptions are metered. According to The New Stack, Anthropic restricted third-party tools from drawing on subscription limits in April, when Claude Code lead Boris Cherny said subscriptions “weren’t built for the usage patterns of these third-party tools.” On May 13 it said that from June 15, automated Agent SDK usage would move onto a separate monthly credit, from $20 on Pro up to $200 on Max. On the day it was due to take effect, Anthropic emailed subscribers that it was pausing the change and that “nothing has changed” for now, adding that it was reworking the plan.
The new credits are a different thing. Anthropic’s help center says so in its FAQ: this is not the Agent SDK credit announced in June, and that credit “isn’t available.” Claude Haiku 5.5, released the same day, also stretches the money further: on prompts under 100K tokens it costs $0.10 per million input tokens and $0.50 per million output tokens.
The facts
Figures and terms below come from Anthropic’s platform docs and help center, checked on October 8, 2026.
| Item | Detail |
|---|---|
| Eligible plans | Max 5x, Max 20x, Team (including discounted Team plans); Free, Pro and Enterprise are excluded |
| Monthly amount | Max 5x $100; Max 20x $200; Team $20 per Standard seat and $100 per Premium seat, pooled, capped at $500 |
| Worked example | Anthropic’s own: 3 Standard + 2 Premium seats = $260 a month; add one Premium seat and the next cycle is $360 |
| Covered | Claude API (including Message Batches), Claude Managed Agents, Claude Agent SDK, Playground |
| Not covered | Interactive Claude Code in the terminal, IDE, desktop or web; extra usage in Claude, Claude Code or Cowork; Bedrock, Vertex AI, Foundry |
| How to claim | Link a Claude Console organization in claude.ai billing settings; Max: the subscriber; Team: an Owner or Primary Owner; new subscribers after 7 days; no payment method needed on Claude Platform |
| Expiry | Granted each billing cycle (monthly on annual plans); unused credits lapse, no rollover |
| Rollout | Anthropic says it rolls out over a few days |
A few rules are easy to miss:
- One plan links to one Console organization, and one organization can take credits from only one plan. You can’t change the link yourself; you have to contact support.
- Included credits are spent before any credits you bought, and auto-reload looks only at your purchased balance. With no credits left, API requests stop until the next cycle. The docs say usage is never charged to your plan.
- Every API key and workspace in the linked organization draws from one balance. To cap a project, give it its own workspace with a spend limit.
claude -pand the Agent SDK use credits when you run them with an API key from the linked organization. Signed in with your Claude plan, they still draw on plan limits. Runs started by the Claude Code GitHub Action, an IDE extension or the desktop app count as Claude Code, so credits don’t apply, even with-p.- The ClaudeDevs account on X says the credits work on any model, “in your code or third-party harnesses.” Read with the docs, that means any call made with the linked organization’s API key.
What others say
Simon Willison called the scheme “really generous” in his October 7 post, noting that the credits exactly match the subscription’s price and make it easier for subscribers to use the API. He also pointed out that OpenAI still lets Codex subscribers use their plan for personal API use, which works out better for heavy API users, and said the new credits go “at least some way” toward closing that gap.
TestingCatalog covered the credits alongside the Haiku 5.5 launch and confirmed the $100, $200 and $500 figures against Anthropic’s help article. The New Stack’s report on the June pause adds context: Zed’s head of growth cited analysis that subscriptions had subsidized third-party usage at roughly 15 to 30 times the equivalent API cost, which helps explain why Anthropic keeps trying to meter automated usage separately. None of these outlets says whether the new credits replace the paused June plan; Anthropic’s docs only say they are separate.
Our take
For developers already on Max, this is real money. It doesn’t touch your plan’s usage limits and needs no new card; what you get is an API-only budget. As arithmetic on list prices, $100 buys about a billion Haiku 5.5 input tokens or about 200 million output tokens (prompts under 100K tokens, before caching and batch discounts). That is a lot of summaries, classification or database-query calls for a month.
The limits are just as clear. The credits can’t be used for the most common subscriber workload, interactive Claude Code; they’re for people building their own things. They reset monthly, so they suit steady small jobs, not stockpiling. Linking the wrong Console organization means a support ticket. Team admins should note that the balance is shared by the whole organization, so without per-workspace spend limits one runaway script can drain it.
What’s still unclear is the subscription-login route for third-party tools: the docs say Agent SDK usage signed in with a Claude account still draws on plan limits, and they say nothing on whether the April restrictions ease. Pro users get nothing here, a tiering worth watching. For how subscription prices compare with API spend, see our piece on Claude plan value versus OpenAI’s.
How to try it
Confirm you’re on Max or Team, then open Settings > Billing on claude.ai (Organization settings > Billing for Team), find API credits, choose the Console organization that should receive them and accept the terms. The credits show up in the Console under Settings > Billing as promotional credits; create an API key and call the API. Anthropic’s docs also suggest the Message Batches API for work that doesn’t need an instant reply, since it costs 50% less and stretches the credits further.