AI Tools
Claude Max API Credits: How to Claim Them, Plus Haiku 5.5 Pricing
In short: since October 7, 2026, a Claude Max or Team subscription includes monthly credits for Anthropic's developer platform: $100 for Max 5x, $200 for Max 20x, up to $500 pooled for a team. They pay for the API, Managed Agents, the Agent SDK (run with an API key) and the Playground, not interactive Claude Code. They do not roll over, and only one organization can receive them: choose it carefully.
The same day, Anthropic released Claude Haiku 5.5, its small model, ten times cheaper per token than Haiku 4.5 as long as the prompt stays at or under 100,000 tokens. The two go together: free credits, and the model that makes them last longest.
I am on Max and claimed mine: $100 landed in the linked organization, without adding a card. Here is everything to know before you do the same, based on Anthropic's official help page.
How Much, and for Which Plan
- Max 5x: $100 a month
- Max 20x: $200 a month
- Team, Standard seat: $20 per seat per month
- Team, Premium seat: $100 per seat per month
- Discounted Team plans (Nonprofit, Scientists): same amounts per seat
On Team, credits for all seats are pooled into one monthly balance, capped at $500. Anthropic's example: a team with three Standard seats and two Premium seats receives $260 a month. The pool is calculated from the seats on the plan when credits are granted.
Free, Pro and Enterprise plans are not eligible.
Eligibility
- an active Max or Team subscription in good standing;
- seven days on an eligible plan: new subscribers wait a week;
- subscriptions bought on iOS or Android count too, but you claim on claude.ai in a web browser;
- no card needed on the platform to claim or use the credit.
Anthropic is rolling the offer out over a few days. If you do not see it yet, wait a little; after seven days on the plan, check that you are on Max or Team with the right role, then contact support.
How to Claim Them
You link a Claude Console organization (the developer platform) to your plan. Credits then arrive there each month.
Before you begin:
- on Max, you must be the subscriber; on Team, a Primary Owner or Owner;
- in the Console, you need the Owner, Admin or Billing role. If you have no organization yet, you can create one during the claim or at platform.claude.com.
The steps:
- On claude.ai, open Settings > Billing (Max subscriber) or Organization settings > Billing (Team Owner or Primary Owner).
- In the API credits section, select Link organization.
- Choose the Console organization that will receive the credits, or create one.
- Review the Supplemental Credit Terms, then select Link organization again.
The credits appear in that organization's balance, ready to use (for me, right away). If you see errors during linking, Anthropic advises trying again in a few hours.
The main trap: you can link only one organization, each organization can receive credits from only one plan, and you cannot change the link yourself: you have to contact support. Link the organization where you actually build. Anthropic also advises signing in to the Console with the same email you use on claude.ai.
What They Cover, and What They Don't
The credits work with any Claude model available on the platform, and cover:
- the Claude API (Messages API and Message Batches API);
- the Playground in the Console;
- Claude Managed Agents;
- the Claude Agent SDK, run with an API key from the linked organization.
They do not cover:
- interactive Claude Code, in the terminal, IDE, desktop app or web;
- extra usage beyond your limits in Claude, Claude Code or Claude Cowork;
- Claude on Amazon Bedrock, Google Cloud Vertex AI or Microsoft Foundry.
The claude -p case. Claude Code's non-interactive command is covered only when you run it yourself with an API key from the linked organization: it is then billed as Agent SDK usage. Signed in with your plan, both claude -p and the Agent SDK keep drawing on your plan's limits. And runs started by the Claude Code GitHub Action, an IDE extension or the desktop app count as Claude Code usage, so they are not covered, even with -p.
How They Work
- Each billing cycle: credits arrive shortly after your plan payment goes through. On annual plans, they still arrive monthly.
- No rollover: unused credits expire at the end of each billing cycle.
- Used first: before any credits you purchased.
- Shared: everyone with an API key in the linked organization draws from the same balance. On Team, workspace spend limits set in the Console limit usage by project or teammate.
- Separate from your limits: they do not change your usage limits in Claude, Claude Code or Cowork.
- Visible in the Console: Settings, then Billing, with amount and expiry date.
When they run out: if the organization has purchased credits or auto-reload, usage continues on those. Otherwise, requests stop until the next grant. Nothing is ever charged to your Claude plan. If your organization is invoiced through Anthropic sales, usage beyond the credits is billed as usual.
If your plan changes: cancelling, downgrading to an ineligible plan or a refund stops new credits, and credits you already have stay usable until they expire. Upgrading from Max 5x to Max 20x gives prorated credits right away, then $200 each cycle. Moving from Max to Team ends the Max link; a Team Owner or Primary Owner can claim the team's credits once the plan has been active for seven days.
Claude Haiku 5.5 Pricing and What to Build
Haiku 5.5 is the model that makes these credits last longest. Its pricing on Anthropic's platform:
- prompts up to 100,000 tokens: $0.10 per million input tokens, $0.50 output;
- above that: $0.50 and $2.50, applied to the whole prompt and its response;
- cache: a read costs $0.01 per million under the threshold;
- batch processing: half price, so $0.05 and $0.25 under the threshold.
For comparison, Haiku 4.5 costs $1 and $5. Above the threshold, Haiku 5.5 is still half the price of Haiku 4.5 per token, but five times its own price under the threshold. For the same text the gap is a bit smaller: according to Anthropic, models from Claude 4.7 onward split the same text into about 30% more tokens. Since the 4.6 generation, it is the only Claude model whose price depends on prompt length: Opus 5.5 and Sonnet 5.5 keep the same rate up to a million tokens.
A rough order of magnitude: a call with 10,000 input tokens and 1,000 output tokens costs $0.0015 under the threshold, without caching. The $100 of a Max 5x plan therefore pays for about 66,000 of them a month.
What Anthropic positions it for:
- high-volume text tasks: classification, summarization, extraction, generation;
- real-time experiences: chat, voice, live support;
- subagents: a stronger model such as Opus plans, Haiku runs well-defined subtasks;
- browser and desktop automation on repetitive tasks;
- simple coding: targeted edits across many files.
It is also the first Haiku with effort controls, per task, to trade cost against intelligence.
My Tips to Get the Most Out of Them
- Link the right organization the first time. The one with your production keys or personal projects, not a test organization.
- Keep each prompt sent to Haiku under 100,000 tokens. A subagent that receives only the relevant passage stays under the threshold; one that receives the whole history can cross it.
- Turn on caching for instructions and tools repeated across calls.
- Batch anything that can wait: half price.
- Do not count on them for Claude Code. Your daily Claude Code usage stays on your plan's limits.
- Spend them within the month. They do not roll over: it is the right time to test an agent or automation idea you kept postponing.
Frequently Asked Questions
Do API credits cover Claude Code?
Not in interactive mode. They are for building your own apps and agents on the platform. The claude -p command and the Agent SDK are covered only when run with an API key from the linked organization.
I am on Pro, am I eligible?
No. Only Max and Team plans, including discounted Team plans, receive these credits.
Can I split the credits across several organizations?
No. They go to a single Console organization, and changing the link goes through support.
What happens when the credits run out?
If the organization has purchased credits or auto-reload, usage continues on those. Otherwise, requests stop until your next monthly credits arrive. Nothing is charged to your Claude plan.
Do unused credits roll over?
No. They expire at the end of each billing cycle.
Is this the Agent SDK credit announced in June?
No. Anthropic states that credit is not available. API credits cover the API, Managed Agents and the Agent SDK, and you claim them into a Console organization.
Up to date as of October 8, 2026. Sources: @claudeai and @ClaudeDevs announcements on X (October 7, 2026), Anthropic's official help page "Monthly API credits for Max and Team plans" (support.claude.com), the Claude Haiku 5.5 page (anthropic.com/claude/haiku) and the platform pricing page (platform.claude.com/docs/en/about-claude/pricing).
To compare the coding agents themselves, also read my Codex vs Claude Code comparison.
Want to turn these credits into a real agent or automation for your business? Book your free audit: we review your project together in 30 minutes.