Anthropic Promised API Credits to Every Paid Plan. It Delivered Them to Two.
Anthropic promised monthly credits to Pro, Max, Team and Enterprise in May. The October 7 version covers Max and Team only, Team capped at $500 pooled.
Published: October 11, 2026 Impact: High. Affects Claude Pro, Max, Team and Enterprise subscribers, and anyone who read the May announcement as a promise.
On May 13, 2026, Anthropic announced a monthly credit for programmatic use on all paid Claude plans, Pro through Enterprise, starting June 15. On June 15 it paused the plan. On October 7 it launched "monthly API credits" instead. Pro and Enterprise get nothing, Team is capped at $500 a month for the whole team, and no plan can spend the credit on interactive Claude Code.
Max subscribers get the dollar amounts they were promised. They cannot spend them on interactive Claude Code.
What Happened (This Week)
The May 13 announcement came from the @ClaudeDevs account on X: "Starting June 15, paid Claude plans can claim a dedicated monthly credit for programmatic usage." It covered the Agent SDK, claude -p, Claude Code GitHub Actions and third-party apps built on the Agent SDK. The post drew 11 million views.
Anthropic's help article, as the Wayback Machine captured it on June 4, said the credit was "available on Pro, Max, Team, and Enterprise plans starting on June 15, 2026." It also said Agent SDK and claude -p usage would stop counting toward plan limits from that date.
The live version of the same article now carries two update lines. June 15: "We've paused the previously-announced changes to Claude Agent SDK usage. For now, nothing has changed." October 7: "Claude Max and Team plans now include monthly API credits."
The October change was announced in Anthropic's October 7 Claude Haiku 5.5 launch post, under "Further updates": "this week, we'll roll out a new monthly API credit to all Max and Team subscribers for use on the Claude Platform. Max 5x users will get $100 in credits per month, Max 20x users will get $200, and Team subscribers will receive up to $500, pooled across their users." The post does not mention Pro, Enterprise, the May promise, or that the credit cannot be used for Claude Code.
May Promise vs October Delivery
| Promised May 13 (for June 15) | Delivered October 7 | |
|---|---|---|
| Pro | $20/month | $0 |
| Max 5x / 20x | $100 / $200 | $100 / $200 |
| Team | $20 or $100 per seat, per user | Same per seat, pooled, capped at $500 per team |
| Enterprise | $20 (usage-based), $200 (Premium seats) | $0 |
Covers signed-in Agent SDK, claude -p, GitHub Actions, third-party apps |
Yes | No, they stay on plan limits |
| Covers interactive Claude Code | No | No |
| Needs a Console (developer) account, linked once | No, one-time opt-in | Yes, and you cannot relink it yourself |
| Rolls over | No | No |
The arithmetic: a Team with 10 Premium seats would have received 10 x $100 = $1,000 a month under the May plan. Under the October plan it receives $500.
What This Means
Three things in the October terms matter to a buyer paying $100 to $200 a month. The credits "don't apply to" interactive Claude Code in the terminal, IDE, desktop or web, or to extra usage. They are spendable on the Claude API, Message Batches, the Console Playground, Managed Agents, and the Agent SDK when run with an API key from the linked Console organization. Unused credits "expire at the end of each billing cycle." And when they run out with nothing else on the account, "API requests stop until your next monthly credits arrive."
Fairness points. Max subscribers receive the amounts announced in May. Nothing here is a price increase. The June plan would also have moved Agent SDK usage off subscription limits. That did not happen, so third-party apps and claude -p signed in with a plan still draw from plan limits, as Anthropic's FAQ states.
The pricing page lists Max with "$100 to $200/month in API credits to build on the Claude Platform" and Team with "Up to $500/month." Both lines are accurate. Neither tells a Claude Code user that the credit is not for Claude Code.
OneHuman did not cover the May 13 announcement when it was made, and its knowledge base holds no Agent SDK events from that period. The record above rests on the Wayback capture and Anthropic's own update lines, not on OneHuman tracking it live. What OneHuman did record is a pattern with Claude plan benefits. BNA #23 documented a "temporary" Claude Code boost extended four times with no date past August 31. BNA #26 documented a "permanent" 25% increase that netted out to a 17% cut from the boosted limit. The cases differ. In each, the headline number and the delivered number were not the same, and the detail sat outside the announcement.
a16z's October 5 consumer AI report puts 7.3% of Claude's US consumer payers on Max, the plan this credit is pitched to.
Same Post, Same Pattern: Haiku 5.5's 100K Price Step
The post that announced the credits also launched Claude Haiku 5.5, which Anthropic says "costs around 75% less to run" than Haiku 4.5. Its price table steps up at 100,000 tokens per prompt (per 1M tokens):
| Up to 100K | Over 100K | |
|---|---|---|
| Input | $0.10 | $0.50 |
| Output | $0.50 | $2.50 |
| Cache reads | $0.01 | $0.05 |
| Cache writes | $0.125 | $0.625 |
Every line is 5x higher above 100K. Anthropic's footnote 2 says Haiku 5.5 is "priced 90% lower than Claude Haiku 4.5 for requests up to 100,000 tokens, and 50% lower for requests over 100,000 tokens. On Haiku 4.5, 90% of requests fell into the former category." The 75% headline is therefore an average weighted by old Haiku traffic, not a price on every task. The post recommends Haiku 5.5 for "compactions", subagent work and browser use, and compaction can easily exceed 100K tokens.
Fairness: above 100K, Haiku 5.5 still costs half of Haiku 4.5 ($0.50 against $1.00 input). It is not a price increase. The same post also halved Sonnet 5.5's cache-read price, from $0.20 to $0.10 per 1M.
Consumer Protection Q&A
Q: I'm on Max 20x. Can I use the $200 for Claude Code when I hit my weekly limit? A: No. Interactive Claude Code in the terminal, IDE, desktop or web is excluded, as is extra usage.
Q: I'm on Pro. Where is my $20? A: Not available. Anthropic's article states that "Free, Pro, and Enterprise plans aren't eligible."
Q: I use Claude in a third-party app with my subscription. Does the credit help? A: No. That usage still draws from plan limits.
Q: Is the money real? A: Yes, if you build on the API with a key from your linked Console organization. It expires each billing cycle, and you must wait seven days on the plan before claiming.
Q: We are a 12-person Team. What do we get? A: Per-seat amounts pool into one monthly balance, capped at $500 for the whole team.
Q: Is Haiku 5.5 really 75% cheaper? A: Below 100K tokens per prompt it is 90% cheaper than Haiku 4.5. Above 100K it is 50% cheaper, and 5x its own sub-100K price. The 75% is an average weighted by how Haiku 4.5 was used.
What Happens Next
- 30 days: Does Anthropic restate the June credit anywhere, or does "That credit isn't available" stand? OneHuman's October 15 Compare Page update will add the credit terms to the Claude column.
- 90 days: Do Pro and Enterprise get any credit, or does the pause become permanent?
- 6-12 months: Does programmatic use ever move off plan limits, as the June plan said it would?
What to Do Next
- If you are on Max for Claude Code, treat the credit as a separate API budget. It will not extend your coding sessions.
- If you build on the API, link your Console organization carefully. You cannot change it yourself afterward.
- If you budgeted for Pro or Enterprise credits from the May announcement, remove them.
- If you run a Team larger than five Premium seats, compare $500 pooled against your seat count before assuming per-seat value.
Bottom Line
What's real: Max and Team receive monthly API credits, at Max amounts matching the May promise. What to watch: whether Pro and Enterprise ever get a credit, and whether Agent SDK usage ever leaves plan limits. What's next: compare the Claude column on the Compare Page against what you actually use before renewing at $100 or $200.
Sources
- @ClaudeDevs on X, May 13, 2026: announcement of the monthly programmatic credit from June 15
- Anthropic Help Center, "Use the Claude Agent SDK with your Claude plan," Wayback capture of June 4, 2026: plans, amounts, per-user terms
- Anthropic Help Center, same article, live version with June 15 and October 7 updates
- Anthropic Help Center, "Monthly API credits for Max and Team plans": October terms, exclusions, FAQ
- Anthropic, "Claude Haiku 5.5," October 7, 2026: announces the Max and Team credit; Haiku 5.5 pricing and footnotes
- a16z, "The Top 100 Gen AI Consumer Apps, 7th Edition," October 5, 2026: 7.3% of Claude's consumer payers on Max
- Anthropic, Claude pricing page, checked October 11, 2026
- OneHuman, BNA #23, August 27, 2026, and BNA #26, September 12, 2026
Verified by OneHuman · October 11, 2026
Independent. AI-assisted. Human-verified. No ads. No affiliates. No investors.
Share This Article
"Anthropic announced a monthly credit for every paid Claude plan in May. The October version covers Max and Team. Pro and Enterprise get $0, and Team is capped at $500 for the whole team."
"A Team with 10 Premium seats would have received 10 x $100 = $1,000 a month under the May plan. Under the October plan it gets $500, pooled for the whole team."
"Max 20x subscribers get the $200 that was promised. The credit still cannot be spent on interactive Claude Code."
"Anthropic's own FAQ asks whether this is the Agent SDK credit announced in June. Its answer, in full: 'No. That credit isn't available.'"