The API is cheaper than Claude Pro below 95 Sonnet messages a day
Claude Sonnet 5.5 costs $2 per million input tokens and $10 per million output, and Claude Opus 5.5 costs $4 and $20, per Anthropic's pricing docs read on 31 October 2026. We priced one typical message at 1,000 tokens in and 500 out. That's $0.007 on Sonnet 5.5 and $0.014 on Opus 5.5. Claude Pro is $20 a month. So $20 buys 2,857 Sonnet messages or 1,429 Opus ones. Over a 30 day month, that's 95 or 48 a day.
Haiku 5.5 never even gets close to the line. At $0.10 in and $0.50 out, it costs $0.00035 a message, and 120 a day for a month comes to $1.26. Below the Pro line, the API is cheaper. Above it, Pro is cheaper only if its limits let you send that many, and Anthropic doesn't say whether they do. The same per message rates for 100 models are in the AI model cost index.
How many messages make Pro, Max or Ultra cheaper than the API?
Claude Pro beats the API at 2,857 Sonnet 5.5 messages a month, Max at 14,286 and Ultra at 28,571, before counting the API credit that comes with Max and Ultra. Each cell is the plan price divided by the API cost of one message of 1,000 tokens in and 500 out, at Anthropic's list prices on 31 October 2026:
| Model | API cost per message | Pro, $20 a month | Pro, $200 a year | Max, $100 a month | Ultra, $200 a month |
|---|---|---|---|---|---|
| Haiku 5.5 | $0.00035 | 57,143 | 47,619 | 285,714 | 571,429 |
| Sonnet 5.5 | $0.007 | 2,857 | 2,381 | 14,286 | 28,571 |
| Opus 5.5 | $0.014 | 1,429 | 1,190 | 7,143 | 14,286 |
| Fable 5.1 | $0.035 | Billed at API rates on Pro | Billed at API rates on Pro | 2,857 | 5,714 |
Only the prices in that table are Anthropic's. The message size, the 30 day month and the idea that a plan's limits stretch that far are our assumptions. Here's what Anthropic does publish about each plan, from its pricing page and help center on 31 October 2026:
| Plan | Price | Allowance Anthropic publishes | Claude API credit included |
|---|---|---|---|
| Free | $0 | Limited usage every 5 hours and every week | None |
| Pro | $20 a month, or $200 a year | More usage than Free, every 5 hours and every week | None |
| Max | $100 a month | 5x Pro, every 5 hours and every week | $100 a month |
| Ultra | $200 a month | 10x Pro every week, no 5-hour limit for verified accounts | $200 a month |
Anthropic gives no message count for any plan. Its Pro plan article says "The number of messages you can send will vary based on message length, including the length of files you attach". By the published multiples, Pro, Max and Ultra buy the same weekly usage per dollar. You get five times the usage for five times the price, or ten times for ten. Ultra is the plan Anthropic called Max 20x until late October 2026. Its Ultra plan article says "The Max 20x plan is now called the Ultra plan", and existing subscribers moved "with no price change". What people actually hit on Pro is in Claude Pro limits.
Max and Ultra pay their fee back as Claude API credit
Claude Max is where the API versus subscription choice mostly disappears. Max costs $100 a month and includes $100 a month of Claude API credit. Ultra costs $200 and includes $200, per Anthropic's monthly API credits article. If you'd spend $100 a month on the API anyway, Max gives you the same $100 of tokens. On top of that, the chat allowance, which is five times Pro's, costs you nothing extra.
That gives a clean rule against Pro. Pro has no API credit, and its help article says so: "The Pro plan does not include API usage through the Claude Console." A Pro subscriber who spends $80 a month on the API pays $100 in total, the same as Max. At $90 of API use, Pro plus the API costs $110 and Max still costs $100. Ultra ties Max plus the API once your API bill reaches $200, and from there its doubled chat allowance costs nothing extra. Below $200, Max plus the overflow costs less.
There are three conditions to it. You claim the credit by linking a Claude Console organization to your plan. "Unused credits expire at the end of each billing cycle," so a quiet month loses them. And they don't cover "Interactive Claude Code in the terminal, IDE, desktop, or web", which draws on plan limits instead. They do cover the Messages and Batches APIs, the Agent SDK, and claude -p run with a Console key. What else Max adds is in Claude Max plan.
Why long chats and Claude Code break even sooner
A chat resends its history, so a real conversation costs more per message than our standard one. On the API, turn ten of a thread carries the nine turns before it as input. Ten turns of 500 tokens each way send 50,000 input tokens, not 5,000. On Sonnet 5.5, that thread costs $0.15, or $0.015 a message. Pro's break even drops from 95 messages a day to 44. On Opus 5.5 it's $0.30 a thread and 22 a day. Cheaper Opus routes are in cheapest way to use Claude Opus.
Two more things push the line down for you. Anthropic's Claude Code docs say Opus 5.5, Sonnet 5.5 and Haiku 5.5 "always use extended thinking", and "Thinking tokens are billed as output tokens", so a 500 token reply is a floor. Anthropic's newer tokenizer also "produces approximately 30% more tokens for the same text" than Sonnet 4.6 and earlier. Caching pulls the other way. A cache hit on Sonnet 5.5 or Opus 5.5 costs 5% of the input price, so a chat client that caches the thread pays far less to resend it. What are tokens in AI explains the units.
Claude Code is the extreme case. Anthropic's cost docs say "the average cost is around $13 per developer per active day and $150-250 per developer per month" across enterprise deployments, at API rates. That's above Max's $100 and around Ultra's $200. At that average, a daily Claude Code user pays less on Max than on the API, as long as Max's limits hold. The per plan detail is in Claude Code pricing.
What does Claude Pro include that the API does not?
Claude Pro is a product, and the API is a model you call. The gap is everything you'd build or go without:
| Claude Pro includes | The API gives you |
|---|---|
| Chat apps on web, desktop and mobile | An endpoint; you bring a chat client |
| Memory across conversations | Nothing kept between requests |
| Projects, Claude in Chrome, and Claude Code on your plan login | Claude Code billed per token with an API key |
| A bill that stops at $20 unless you turn on usage credits | A bill that grows with use until you set a spend limit in the Console |
The API wins when what you want is control over it all. You pick the model for each request, and you can send anything that can wait through the Batch API at 50% off. Claude 4.6 and later models, except Haiku 5.5, get the full 1M token context at standard rates. New accounts get "a small amount of free credits to test the API", per the pricing docs. The OpenAI side of this choice is in ChatGPT API vs subscription, and the two price lists side by side are in ChatGPT API vs Claude API pricing.
Pick the API under 95 messages a day, Pro above it, Max past $80 of API use
Use the API if you send fewer than about 95 Sonnet 5.5 messages a day in short threads, or 44 in long ones, and you're happy to run a chat client. Set a spend limit before the first request.
Use Pro if you chat or code with Claude most days and want the apps, memory and Projects. It's $20 or $200 a year. Full plan prices are in Claude pricing.
Use Max if you also build on the API. Its credit covers the first $100 of API use, so it costs less than Pro plus the API once your API bill passes $80.
Whizi, which publishes this page, is a fourth route, and it doesn't win on Claude alone. It has no Claude API and no Claude Code. Claude Sonnet 5.5 costs 10 credits a message there and Opus 5.5 costs 20, from 2,500 credits on Pro at $29.99. Per message, that's far more than the API's $0.007. It fits someone who wants Claude, ChatGPT and Gemini in one app under one capped bill. The savings calculator adds up what your separate subscriptions cost now.
- Count your Claude messages a day and how long your threads run before choosing
- Price a month at $0.007 per Sonnet 5.5 message and $0.014 per Opus 5.5 message, then double it for long threads
- Set a spend limit in the Claude Console before the first API request
- If you pay for Pro and spend more than $80 a month on the API, switch to Max and claim its $100 credit
- Claim Max or Ultra API credit each month; unused credit expires at the end of the billing cycle
Frequently asked questions
Is the Claude API cheaper than Claude Pro?
Yes, below about 95 Sonnet 5.5 messages a day. At Anthropic's list prices on 31 October 2026, a message of 1,000 tokens in and 500 out costs $0.007 on Sonnet 5.5 and $0.014 on Opus 5.5, so Pro's $20 equals 2,857 Sonnet or 1,429 Opus messages a month. Long threads cut those numbers roughly in half.
Does Claude Pro include API access?
No. Anthropic's Pro plan article says "The Pro plan does not include API usage through the Claude Console." The API is billed separately per token. Max includes $100 a month of API credit and Ultra $200, and unused credit expires at the end of each billing cycle.
Can I use Claude Max API credits in Claude Code?
Only outside interactive sessions. Anthropic's credits article says they cover claude -p and the Agent SDK run with a Console API key, but not "Interactive Claude Code in the terminal, IDE, desktop, or web". Interactive sessions draw on your Max plan limits instead.
How many messages do you get on Claude Pro?
Anthropic doesn't publish a number. Its pricing page says Pro gives "More usage than Free" every 5 hours and every week, and its help center says the count varies with message length and attachments. Max gives 5x Pro and Ultra gives 10x Pro a week. Claude Pro limits collects what users hit in practice.
