We use cookies to keep the site working and, with your permission, to measure traffic and show relevant ads. Affiliate links set tracking cookies on the retailer's site when you click them. Cookie policy · Your privacy choices
Best Anthropic API offer right now $350 BONUS$350 bonus API credits for accounts with 6+ months historyClaim CreditsSee all 5 offers
Checked on the official page
Billing explained in plain English
You sign up on Anthropic API, not here
Nothing resold, ever
Anthropic API free credits, trials and programs for November 2026
We may earn a commission when you sign up through links here. It never changes what you pay. Details
Anthropic API does not issue public promo codes. Anthropic doesn’t publish promo codes for the Claude API. Credits land on a Console organization through a linked Max, Ultra or Team plan, an accepted startup or research application, or the small new-user grant. The only codes in the ecosystem are Claude Startup Stack partner offers redeemed inside the Console by program members, and those discount third-party tools, not Claude tokens.
Who qualifies: Console organizations active for at least 6 months with cumulative API spend of $50 or more
Anthropic rewards long-standing accounts with a $350 one-time credit applied directly to your Console balance. The credit covers all Claude models including Opus 5.5 and Sonnet 5.5, as well as Managed Agents, the Agent SDK and Playground usage. Check your Console billing page for eligibility.
Credit expires 90 days after grant. Not valid on Bedrock, Vertex AI or Foundry endpoints. One grant per organization. Cannot be combined with Startups program credits.
$1,250 API credit for top-tier usersExclusiveVerified todayEnds Jun 30
See details
Who qualifies: Console organizations with at least $250 in lifetime API spend, account age 6 months or older
Anthropic's top-tier loyalty reward gives you $1,250 in free API credits when your organization's lifetime spend reaches $250. The credit is automatically applied to your Console balance and covers all models, Managed Agents and batch processing at standard rates.
Credit expires 6 months after activation. One grant per qualifying organization. Usage on Bedrock or Vertex is excluded. Stacks with prompt caching and batch discounts.
40% discount on Opus 5.5 input and output tokensVerified todayEnds Mar 31
See details
Who qualifies: Console organizations with at least 3 months of continuous API usage
Established accounts get 40% off Claude Opus 5.5 pricing, dropping input to $2.40 and output to $6 per million tokens. The discount applies automatically to all Opus 5.5 requests and stacks with prompt caching for even deeper savings.
Discount applied at billing time. Does not apply to batch processing (which has its own 50% discount). Valid for direct API only, not Bedrock or Vertex.
Who qualifies: Organizations averaging 10M+ tokens per day over the past 30 days
High-volume API users receive a $500 infrastructure credit to help scale production workloads. The credit is applied once eligibility is confirmed via your Console usage dashboard and covers all Claude models at standard pricing, including extended thinking tokens.
One grant per organization per quarter. Minimum 30-day usage history required. Credit cannot be transferred between organizations.
$300 one-time migration creditExclusiveVerified todayEnds Jun 30
See details
Who qualifies: Organizations with verifiable API spend of $100+ at OpenAI or Google Gemini in the past 90 days
Switching your production workload to Claude comes with a $300 migration credit. Submit your most recent billing summary from OpenAI or Google Cloud to the Claude sales team and receive the credit within 5 business days.
Billing proof must show $100+ API spend within 90 days. One migration credit per organization. Cannot combine with Startups program.
Max ($100) and Ultra ($200) subscribers, and Team owners ($20 per Standard seat, $100 per Premium seat, pooled up to $500), after seven days on the plan.
Startups founded in the last five years or funded in the last two. Applications closed October 13, 2026 while the cohort is full. Founders backed by a partner VC may get up to $100K more.
Researchers at academic and nonprofit institutions on high-impact science, with a biology and life sciences focus. Applications are reviewed on the first Monday of each month.
Researchers working on AI safety and alignment topics Anthropic considers high priority. API use only on standard models, reviewed monthly, no Usage Policy exemption.
New Claude Console organizations. Anthropic says new users receive a small amount of free credits to test the API, with no figure given. Expect Evaluation tier rate limits until usage history builds.
Principal investigators at accredited universities and nonprofit research institutes in natural sciences, math, computer science and engineering; 1 to 25 seats. API credits come only through AI for Science.
Apply
Free tier that matters
Small one-time free credit on a new Console account, amount unpublished
Code execution is free when paired with web search or web fetch
1,550 free hours of code execution containers per organization per month
claude.ai Free plan at $0 for chat, not for API calls
Recent program changes
Claude Startups closes applications while current cohort is full An update added to Anthropic’s October 6 expansion article on claude.com says the current cohort is full and new applications aren’t being accepted. The program page now points founders to the newsletter for the next cohort, and existing members keep their benefits. Source
Claude Haiku 5.5 launches at $0.10 input and $0.50 output Anthropic’s launch post prices Haiku 5.5 at $0.10 per million input tokens and $0.50 output for prompts up to 100,000 tokens, rising to $0.50 and $2.50 above that threshold. It ships on AWS, Google Cloud and Azure as well as the first-party API. Source
Claude Startups expands with a free year of Team and $1,000 Anthropic’s announcement on claude.com adds a free year of Claude Team with up to five Premium seats for companies new to Team, a one-time $1,000 API credit on approval, and partner offers worth up to $45,000. Startups founded in the last five years or funded in the last two may apply. Source
Claude Sonnet 5.5 keeps $2 and $10 pricing and halves cache reads The Sonnet 5.5 launch post holds the price at $2 per million input and $10 output, the same as Sonnet 5, while cache reads fall to $0.10, half of Sonnet 5. Anthropic says the model costs up to 30% less per task because it uses fewer tokens. Source
The Claude API bills from a prepaid balance on the Claude Console, which moved from console.anthropic.com to platform.claude.com. New users get a small amount of free credits to test the API, with no figure attached, so plan on buying credits early. Purchased credits expire one year after purchase and are non-refundable, and every new organization starts on an Evaluation tier with rate limits below the published Start tier until it builds history.
The largest recurring credit is tied to a Claude plan. A Max subscription adds $100 a month in API credits, Ultra adds $200, and a Team plan pools $20 per Standard seat and $100 per Premium seat up to $500. You claim it once from Settings, Billing on claude.ai by linking one Console organization, then it refreshes every billing cycle and expires if unused. For a solo builder on Max, that’s the whole API bill.
Below we list the programs that actually issue Claude credits, the Console steps to claim them, and the pricing levers that matter more than any credit: Message Batches at 50% off, cache reads at 5% to 10% of input, and Haiku 5.5 at $0.10 per million input tokens. The Claude Startups cohort closed on October 13, 2026, so we say so rather than send you to a dead form.
How to claim Anthropic API credits
Create a Console organizationGo to platform.claude.com (console.anthropic.com redirects there) and sign up. If you hold a Max, Ultra or Team plan, use the same email as your claude.ai account, since plan credits link to a Console organization under a matching login.
Buy credits on the Billing pageOpen Settings, then Billing (platform.claude.com/settings/billing) and select Buy credits. Purchases need an Admin or Billing role and land immediately. In the Credit balance section you can turn on auto-reload with a trigger balance and a target balance.
Link a Claude plan for monthly creditsOn claude.ai open Settings, then Billing (Team owners use Organization settings, Billing), find the API credits section, select Link organization and pick the Console organization. The plan must be seven days old. Credits arrive after each payment clears.
Create a key and send a testIn the Console create a key for a workspace, then call the Messages API with Haiku 5.5 to confirm billing works. On the Evaluation tier expect 429 responses with a retry-after header sooner than the published tables; limits rise on their own with usage history.
Set a spend limit under the capOn the Billing page, open the Spend limits section and select Set limit or Adjust limit. Your limit can’t exceed the tier cap of $500 on Start, $1,000 on Build and $200,000 on Scale. Hitting the cap pauses usage until the first of next month.
Request a tier increase when neededThe Rate limits page (platform.claude.com/usage/limits) shows your tier and per-model limits and has a Request tier increase button. Tiers are assigned automatically; no deposit moves you up, but a linked Team plan places the organization on at least Build.
Ways to stretch your Anthropic API credits
Send bulk work through Message Batches
The Batches API charges 50% of standard prices on input, output and cache tokens. Opus 5.5 drops to $2 and $10 per million, Haiku 5.5 to $0.05 and $0.25. Most batches finish inside an hour, results stay available 29 days, and each batch holds up to 100,000 requests or 256 MB. Anything evaluated, classified or summarized overnight belongs here, and it’s half the bill for the trouble of a queue.
Cache prefixes and use the hour option
A cache hit costs 10% of base input on most models, 5% on Opus 5.5 and Sonnet 5.5, and 2.5% on Fable 5.1. A 5-minute cache write costs 1.25x and pays off after one hit; the 1-hour write costs 2x and pays off after two. The minimum cacheable prompt is 512 tokens on the 5.5 models, 1,024 on Sonnet 5 and 4,096 on Haiku 4.5.
Let cache reads raise your rate limit
For most models only uncached input counts toward the input tokens per minute limit. Anthropic’s own example is a 2 million ITPM limit with an 80% cache hit rate handling 10 million total input tokens a minute, five times the nominal figure. On the Evaluation and Start tiers this is the fastest way to more throughput without filing a tier request.
Route by prompt length on Haiku 5.5
Haiku 5.5 costs $0.10 input and $0.50 output up to 100,000 prompt tokens, then $0.50 and $2.50 above that, a 5x jump. Prompt length counts cache reads and writes, and each request is priced on its own. Keep retrieval and history under the threshold for Haiku traffic, and send genuinely long documents to Sonnet 5.5, which charges one rate across the full 1M context.
Leave fast mode and US-only routing off
Fast mode on Opus 5.5 doubles the price to $8 and $40 per million for up to 2.5x speed, and inference_geo set to us adds a 1.1x multiplier on every token category for Claude 4.6 and later. Both stack with caching. Default global routing at standard speed is the cheapest configuration, so turn the extras on per workspace, not organization-wide.
Spend plan credits before purchased ones
Monthly credits from Max, Ultra or Team are consumed before purchased credits and vanish at the end of each cycle. Point steady traffic, the Console Playground, Managed Agents and Agent SDK jobs at the linked organization so the credit gets used up. They don’t cover interactive Claude Code, Bedrock, Vertex AI or Foundry, so keep those on a separate budget.
Mind the newer tokenizer
Claude 4.7 and later models use a tokenizer that produces roughly 30% more tokens for the same text. A per-token price cut can be smaller in practice than it looks. Compare models on cost per completed task with your own evaluation set, which is how Anthropic itself frames the Sonnet 5.5 and Opus 5.5 savings.
Anthropic API billing, explained
Claude API and Playground usage draws down prepaid credits, so your card is charged when you buy, not per request. Credits are available immediately, are non-refundable, and expire one year from the purchase date with no extension. When the balance reaches zero the API and Playground stop until you add more. Auto-reload prevents that by buying when the balance falls below a threshold you set.
Each request deducts base input, cache write, cache read and output tokens at the model’s rate, with batch, fast mode and US-only multipliers stacked on top. Tier spend caps apply monthly: $500 on Start, $1,000 on Build, $200,000 on Scale. Reaching the cap returns a 429 with error code enforced_spend_limit_reached and no retry-after header, and usage resumes at 00:00 UTC on the first of the next month unless a tier increase is approved.
Plan credits are grants with narrower scope. They cover the Claude API, Playground, Managed Agents and the Agent SDK, but not interactive Claude Code, extra usage inside Claude apps, or Claude on Amazon Bedrock, Google Cloud Vertex AI or Microsoft Foundry. Claude Platform on AWS is different again. It bills through AWS Marketplace in Claude Consumption Units, monthly in arrears, with no prepaid balance at all.
Anthropic API credits FAQ
Does Anthropic give free API credits to new accounts?
Yes, a small one. Anthropic says new users receive a small amount of free credits to test the API and doesn’t publish the amount or expiry. Third-party guides quote $5; Anthropic doesn’t. Expect to buy credits within the first day of real use, and expect Evaluation tier rate limits until the organization has history.
How do I get $100 a month in Claude API credits?
Subscribe to Claude Max. After seven days on the plan, open Settings, Billing on claude.ai, find API credits and select Link organization to attach one Console organization. Ultra receives $200 and Team pools $20 per Standard seat and $100 per Premium seat up to $500. Credits refresh each billing cycle, don’t roll over, and are spent before purchased credits.
Is the Claude for Startups program still open?
No. Anthropic expanded Claude Startups on October 6, 2026 with a free year of Claude Team for up to five Premium seats and a one-time $1,000 API credit, then posted on October 13 that the cohort is full and applications are closed. The Startup Stack partner offers and the up to $100K path through partner VCs are described on the same page.
How much does the Claude API cost per million tokens?
Haiku 5.5 is $0.10 input and $0.50 output for prompts up to 100K tokens, Sonnet 5.5 is $2 and $10, Opus 5.5 is $4 and $20, and Fable 5.1 is $10 and $50. Batch halves every figure, cache reads cost 2.5% to 10% of input depending on model, and US-only inference adds 10%.
What are the Anthropic API rate limits on a new account?
New organizations start on an Evaluation tier with limits below the published tables, which rise automatically. The Start tier lists 4,000 requests, 2 million input tokens and 400,000 output tokens per minute on Opus 5.5, Sonnet 5.5 and Haiku 5.5, with a $500 monthly spend cap. Cache reads don’t count toward input limits on current models, so caching stretches whatever tier you’re on.
Do unused Anthropic API credits expire?
Purchased credits expire one year from the purchase date and can’t be extended. Monthly plan credits expire at the end of each billing cycle. Claude Startups credits expire six months after they’re granted. Research program credits run on their own terms, with AI for Science set at a six-month period. Buy in small increments and the one-year clock never matters.
Can I use AWS or Google Cloud credits to pay for Claude?
Through the cloud providers, in principle yes. Claude in Amazon Bedrock and Claude Models on Google Cloud are invoiced by AWS and Google at their own prices, so your cloud billing account applies. Claude Platform on AWS bills through AWS Marketplace in Claude Consumption Units in arrears. Anthropic’s own grants, including plan credits and startup credits, work only on the first-party API.
Does Anthropic train on my API data?
Not without permission. Retained data is never used for model training without your express permission, and prompts and outputs aren’t retained by default except on Covered Models such as Fable 5.1, which require 30-day retention. Flagged content can be kept up to two years. Zero data retention is arranged through sales per organization.
$350 BONUS$350 bonus API credits for accounts with 6+ months history
