Claude Max 20x is now Ultra: how to use its monthly API credits

What changed on 30 October

Ultra dropped the five-hour limit, the usage window that resets every five hours, for verified accounts. A new Ultra account keeps a temporary one until Anthropic checks it, a guard against fraud and abuse by The Next Web's account, and the subscriber gets an email when it lifts. Max keeps its five-hour window.

How to claim the credits

1. Wait until your plan has been active for seven days, if it is new. The subscriber claims on Max or Ultra, and the Primary Owner or an Owner on a Team plan.
2. On claude.ai, open Settings > Billing, or Organization settings > Billing on a Team plan.
3. In the API credits section, choose Link organization. Pick a Claude Console organization where you are an Owner, Admin or Billing member, or create one in the same flow. The Console side needs no payment card.
4. Read the credit terms and link. Choose the organization you will build in, because a plan links to one organization and only support can change it.
5. In the Console, open Billing. The credits sit under Promotional credits.
6. Create an API key in that organization.

If your settings show no API credits section, the seven days may not have passed, you may lack the role, or the rollout may not have reached you yet. Anthropic says it takes a few days. Contact support if you think you're eligible.

What the credits pay for

Anthropic's docs say the credits don't change your usage limits in Claude or Claude Code. A claude -p script or an Agent SDK app run with an API key from the linked organization spends the credits. Signed in with your Claude account instead, it uses your plan's usage limits.

Point an interactive session at a key from an organization that holds only these credits, and it stops with this message:

Credit balance too low · Add funds: https://platform.claude.com/settings/billing

What two hundred dollars buys on each model

API prices are per million tokens, split into what you send and what the model writes back. Our example is one request of 3,000 input tokens and 1,000 output tokens, about the size of a support reply that quotes some context, priced at the base rates on Anthropic's pricing page.

At that size, Claude Haiku 5.5 runs the request a hundred times for every one on Claude Fable 5.1. On Sonnet, Ultra's credits cover roughly 400 of them a day. Max's credits buy half of each figure.

What to run on the credits

This request, sent with a key from the linked organization, has Haiku label a support ticket:

curl https://api.anthropic.com/v1/messages \
  -H "x-api-key: $ANTHROPIC_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude-haiku-5-5",
    "max_tokens": 1000,
    "messages": [{"role": "user", "content": "Label this ticket as billing, bug or question: my invoice shows two charges this month."}]
  }'

• If Haiku labels a batch of your own tickets correctly, you don't need to pay for a bigger model. The Console playground runs on the same credits, so you can compare two models on the same tickets before you write any code. When a day's tickets can wait until morning, the Message Batches API charges half the price on input and output alike and returns the results later.
• An agent that works through the support queue on its own can run on Claude Managed Agents, still in beta, in a sandbox Anthropic hosts. It needs no card on file, and the credits pay for it.

Make the credits last

Every key in the organization draws on one balance, so one runaway script can empty it. A workspace spend limit caps each project, and the Console's Cost page shows where the credits went.

With prompt caching, a cache hit on Sonnet costs a twentieth of its normal input price. Writing to the five-minute cache costs a quarter more than normal input, so a long system prompt you resend pays off from its second use within those five minutes. The one-hour cache costs twice the input price to write.

Reference: expiry, top-ups and plan changes

Credits arrive shortly after each plan payment clears, or monthly on a yearly plan. Whatever is left expires at the end of that cycle, and the Console's Billing page shows that date beside the amount.

• Included credits are spent before any credits you bought. Auto-reload watches only the purchased balance, so it can top up while included credits remain.
• When the credits run out, requests stop until the next ones arrive, unless the organization holds purchased credits. Auto-reload in Console Billing keeps them going.
• An organization's monthly spend cap resets on the first of the calendar month and counts credit-paid usage, while the credits follow your billing date.
• A linked organization moves up to at least the Start rate-limit tier, and to Build when its monthly credits are over two hundred dollars. Ultra sits exactly on that line, so only a bigger Team pool reaches Build.
• Upgrading from Max to Ultra brings prorated credits at once, then Ultra's full amount each cycle. Through Google Play, the larger credit can arrive as late as your next renewal.
• Cancelling, a refund or a move to a plan without credits stops new credits when the plan ends. Credits you already hold stay usable until they expire.
• Moving to a Team plan ends the link, and a Team owner can claim the team's pool after the usual seven days.

Last checked on 4 November 2026 against Anthropic's pages under Sources. Mindket has not claimed these credits, so this is a reading of the docs, not hands-on use.

Frequently asked questions

Sources

1. Claude: Plans and pricing (read November 4, 2026)https://claude.com/pricing
2. Claude Help Center: Monthly API credits for Max, Ultra, and Team plans (read November 4, 2026)https://support.claude.com/en/articles/17154008-monthly-api-credits-for-max-ultra-and-team-plans
3. Claude API docs: API credits for Max, Ultra, and Team plans (read November 4, 2026)https://platform.claude.com/docs/en/about-claude/api-credits-for-subscribers
4. Claude API docs: Pricing (read November 4, 2026)https://platform.claude.com/docs/en/about-claude/pricing
5. Claude Help Center: Use the Claude Agent SDK with your Claude plan (read November 4, 2026)https://support.claude.com/en/articles/15036540-use-the-claude-agent-sdk-with-your-claude-plan
6. Anthropic: Supplemental credit terms (read November 4, 2026)https://www.anthropic.com/legal/credit-terms
7. The Next Web: Anthropic renames Claude Max plans and drops Ultra's five-hour limit (read November 4, 2026)https://thenextweb.com/news/claude-ultra-max-plan-rename-five-hour-limit