AI engineering

Claude Startups Program: Credits, Eligibility and How to Apply (2026)

What Anthropic says the Claude Startups program offers, who it is for, why applications are closed right now, and how to budget Claude API costs while you wait.

By EpiFive Editorial TeamReviewed/updated 11 min readHow we check articles
On this page

The Claude Startups program is Anthropic's program for early-stage founders building on Claude. Its documented benefits are discounts and credits worth up to $45,000 from companies building with Claude, live sessions with Anthropic's Applied AI team, and event invitations. On 2026-11-19 the official page said the current cohort was full and applications were closed, so the practical answer today is to prepare, subscribe for updates, and build on pay-as-you-go pricing in the meantime.

This guide explains what the program page and the official program terms actually say, which claims you will see elsewhere that Anthropic does not make on its own pages, how to prepare an application that matches the stated review factors, and how to estimate Claude API costs so you are not waiting on credits to start. Everything is dated, because programs like this change quickly.

Who this is for

This is for founders, technical co-founders and small product teams who plan to build a product on the Claude API, and for people who searched for "Claude startup credits" and found conflicting numbers on blogs and aggregator sites. By the end you will know what is documented, what is only reported, what the terms restrict, and what to do while the cohort is closed.

It is not a guide for individuals who only want to use the Claude chat app, and it does not cover other vendors' startup programs. For a cross-vendor view of free credits, see AWS vs Azure vs Google Cloud free tiers compared.

What the program page says

The program page on claude.com describes Claude Startups as "Anthropic's program for early-stage founders building on Claude." These are the points the page states, read on 2026-11-19.

ItemWhat the page says
AudienceEarly-stage founders building on Claude
Partner offersDiscounts and credits worth up to $45,000 from companies building with Claude, through the Claude Startup Stack, redeemed with a code in Claude Console
VC routeFounders backed by a VC in Anthropic's partner network may be eligible for up to $100K in additional API credits, claimed through their VC
CommunityLive Applied AI group sessions every other week, with an AMA, plus invitations to events, hackathons and meetups
Credit scopeOnly the first-party Claude API through Claude Console; not AWS, Google Cloud, Azure or other third-party platforms
Status"Our current cohort is now full, and we're not accepting new applications right now"

Two details deserve attention. First, the $45,000 figure is a ceiling on the combined value of offers from partner companies, not a Claude API credit balance. The page describes those offers as coming from "companies building with Claude." Treat it as a bundle of vendor discounts that you redeem one at a time. Second, the larger $100K API credit figure is tied to a venture investor in Anthropic's partner network. The page says you claim it through your VC, so a founder without such an investor should not count on it.

What happens to pending applications

The page's FAQ says applications still in review were not advanced because the cohort filled. Applicants were emailed and are encouraged to reapply when the next round opens. If you applied earlier and heard nothing, check the email address you used and your spam folder.

How to be notified

The page says to subscribe to the Claude Startup newsletter to learn when applications open for the next cohort. No opening date is published. Current members sign in through a link on the same page.

What the official terms add

Anthropic publishes separate official terms for its startup program. The version read on 2026-11-19 shows "Last updated May 22, 2025", so it may predate the current cohort structure. It is still the binding text for awards under that program, and it contains points the marketing page does not.

  • Discretion. "Program eligibility is determined in Anthropic's sole discretion" (Section 1). A good application is not a guarantee.
  • Review factors. Section 4 lists "business traction, investment, and funding, as well as Claude integration and usage."
  • Completeness. "Only fully complete and compliant applications are eligible to receive Program Benefits" (Section 3).
  • What awardees may receive. "Anthropic API credits, priority rate limits, and/or access to certain resources" (Section 6). The terms do not state credit amounts, expiry dates or permitted uses.
  • Account attribution. Benefits are attributed to the account for the company identified in your application (Section 6). Apply with the company's Claude Console organization, not a personal one you may later abandon.
  • Tax. Section 8 says API credits "may have different tax implications than cash awards," and that credits may be forfeited if required tax information is not provided. Ask your accountant how credits are treated in your jurisdiction.
  • Eligibility exclusions. The terms list countries and regions whose residents, or those who serve persons there, are ineligible, including Belarus, China, Cuba, Iran and Russia. The terms also exclude "employees, officers, directors, agents, representatives and their immediate families"; read Section 2 for who that covers.
  • Changes. Anthropic may cancel, terminate, modify or suspend the program at its discretion (Section 9).

Numbers you may see that are not on Anthropic's pages

Search results for this topic include blogs, newsletters and partner pages that quote other amounts, such as a free year of a team plan, a $1,000 API credit, or a $30,000 cap for a specific accelerator's participants. None of these appeared on the official program page or the official terms when we read them. They may come from older cohorts, partner arrangements or third-party summaries. This article does not repeat them. If a number matters to your plan, ask the source where it came from, and confirm it with Anthropic or the partner program that offers it.

Preparing an application

Because the page is closed, you cannot submit anything today. You can still prepare, using the review factors Anthropic published. The following checklist maps each factor to evidence you can gather now.

Business traction

Keep simple, accurate records: signed pilots, waitlist counts, revenue, active users, or letters of intent. State only numbers you can show. The terms require applications to be complete and compliant, so a vague or inflated entry is a risk.

Investment and funding

Write down your funding status plainly: bootstrapped, accelerator-backed, pre-seed, seed, and who invested. If an investor is a VC in Anthropic's partner network, ask them how credit claims work. The page says those credits are claimed through the VC.

Claude integration and usage

This is the one factor you control fully. Create a Claude Console organization for the company, build the product against the API, and keep usage records. To see what your usage costs, use the worked example below, and read Claude API pricing and how to estimate token cost.

Housekeeping

  • Use the company's name and a company email address consistently across the Console account and the application.
  • Confirm that you and your company are not in an excluded country or region listed in the terms.
  • Subscribe to the newsletter now so you see the reopening announcement.

Budget the API while you wait

Credits are optional for getting started. Claude API usage is billed per million tokens (MTok), and the pricing page lists these base rates for three current models, read on 2026-11-19:

ModelInput per MTokOutput per MTokBatch inputBatch output
Claude Opus 5.5$4$20$2$10
Claude Sonnet 5.5$2$10$1$5
Claude Haiku 5.5 (prompts up to 100,000 tokens)$0.10$0.50$0.05$0.25

The Batch API gives a 50 percent discount on both input and output tokens, and cache hits are billed at a fraction of the base input rate, for example $0.10 per MTok for Sonnet 5.5.

Worked example

Suppose a prototype processes 5 million input tokens and 1 million output tokens in a month. The cost is input MTok times the input rate plus output MTok times the output rate:

Sonnet 5.5:        5 x $2    + 1 x $10   = $10 + $10   = $20.00
Haiku 5.5:         5 x $0.10 + 1 x $0.50 = $0.50 + $0.50 = $1.00
Opus 5.5:          5 x $4    + 1 x $20   = $20 + $20   = $40.00
Sonnet 5.5 batch:  5 x $1    + 1 x $5    = $5 + $5     = $10.00

Use your own token counts, because real usage depends on prompt length and answer length. The point is that prototype-scale usage is small, so a closed cohort does not have to stop you from building. To cut costs further, see prompt caching for cost and latency and, for work that can wait, the Message Batches API in Python.

Limits that matter for a young account

Anthropic's rate limit documentation says new organizations may start in an Evaluation tier with limits below the standard limits, and that organizations move up automatically as they build usage history. The Start, Build and Scale tiers carry monthly spend caps of $500, $1,000 and $200,000 respectively. You can set your own lower limit under Spend limits in Claude Console, next to Monthly spend limit, by choosing Adjust limit or Set limit. Setting a limit early protects a small budget from a runaway loop.

When an organization reaches its tier cap, API usage pauses until 00:00 UTC on the first day of the next month, and requests return HTTP 429 with error.details.error_code set to enforced_spend_limit_reached. If you need more, use Request tier increase on the Rate limits page in Claude Console.

Troubleshooting and common questions

You cannot find an application form. That matches the page status: applications are closed. Do not use third-party "apply here" links for credits. Use only the claude.com program page.

A blog says the program gives X dollars and your quote differs. Compare it with the table above and the dates. Anthropic's current page lists only the $45,000 and $100K figures.

You want to use credits through a cloud marketplace. The page says credits apply only to the first-party API in Claude Console. If your architecture depends on another platform, plan to pay for that usage separately.

Your Console organization was created under a personal account. The terms attribute benefits to the company identified in the application. Create or move to a company organization before you apply.

What to do this week

  1. Subscribe to the Claude Startup newsletter from the program page.
  2. Create a company Claude Console organization and set a monthly spend limit.
  3. Build your prototype on the cheapest model that meets your quality bar and record real token counts.
  4. Collect traction and funding evidence in one document.
  5. Read the official terms before you apply, especially Sections 3, 4, 6 and 8.
  6. Revisit the program page before each planning cycle, since the status, amounts and terms can change.

What we checked

  • Opened the Claude Startups program page on claude.com on 2026-11-19. It said the current cohort was full and applications were closed, and listed the $45,000 partner offers, the $100K VC-referred API credits and the biweekly Applied AI sessions.
  • Cross-checked the program page against the claude.com/startups address; both gave the same status, amounts and FAQ answers, including that credits work only on the first-party Claude API.
  • Opened the official Anthropic Startup Program terms on 2026-11-19, which show "Last updated May 22, 2025", and quoted Sections 1, 3, 4, 6, 8 and 9. The terms state no credit amounts.
  • Opened the Claude rate limits page and the pricing page in the Claude API docs on 2026-11-19 and copied the tier spend caps ($500, $1,000, $200,000) and the Opus 5.5, Sonnet 5.5 and Haiku 5.5 rates.
  • Recalculated the worked example: Sonnet 5.5 is 5 x $2 + 1 x $10 = $20.00; Haiku 5.5 is 5 x $0.10 + 1 x $0.50 = $1.00; Opus 5.5 is 5 x $4 + 1 x $20 = $40.00; Sonnet 5.5 batch is 5 x $1 + 1 x $5 = $10.00.
  • Searched for third-party coverage and found figures that do not appear on Anthropic's own pages; we left them out.
  • Not checked hands-on: we did not apply, redeem a Startup Stack code or use a Console account, and the application form was not available. Steps follow Anthropic's published pages as of 2026-11-19.

Last tested: 2026-11-19

References

Questions people ask

Is the Claude Startups program open for applications?

On 2026-11-19 the official program page said the current cohort was full and Anthropic was not accepting new applications. It gave no date for the next cohort and pointed people to the Claude Startup newsletter for updates. Check the page again before you plan around it.

How much credit does the Claude Startups program give?

Anthropic documents two figures: discounts and credits worth up to $45,000 from companies building with Claude, and, for founders backed by a VC in Anthropic partner network, up to $100K in additional API credits claimed through that VC. These are maximums, not guarantees, and the program terms say eligibility is decided at Anthropic sole discretion.

Can Claude Startups credits be used on AWS, Google Cloud or Azure?

No. The program page states that credits apply only to the first-party Claude API through Claude Console, not to AWS, Google Cloud, Azure or other third-party platforms. If you plan to call Claude through a cloud marketplace, the credits will not cover that usage.

What does Anthropic look at when it reviews an application?

The official program terms list business traction, investment and funding, and Claude integration and usage as factors, and say the list is not limited to those. Only fully complete and compliant applications can receive benefits. The terms also say awarded API credits may have different tax implications than cash.

ClaudeAnthropicClaude ConsoleClaude APIStartup Credits