Claude API Pricing in 2026: What Each Model Really Costs
Claude has four API tiers, and the price gap between them is large. Here are the standard list prices per million tokens as of November 2026.
Current prices (input / output)
| Model | Input $ / 1M | Output $ / 1M | Note |
|---|---|---|---|
| Claude Fable 5.1 | $10 | $50 | |
| Claude Opus 5.5 | $4 | $20 | |
| Claude Sonnet 5.5 | $2 | $10 | |
| Claude Haiku 5.5 | $0.10 | $0.50 | for prompts up to 100,000 tokens, then $0.50 / $2.50 above that |
Batch processing halves these rates, and cached input is far cheaper (see our guide to prompt caching).
What a typical request costs
Take a request with 100,000 input tokens and 20,000 output tokens:
| Option | Cost (approx.) |
|---|---|
| Fable 5.1 | $2.00 |
| Opus 5.5 | $0.80 |
| Sonnet 5.5 | $0.40 |
| Haiku 5.5 | $0.02 |
Haiku is roughly twenty times cheaper than Sonnet on this job, which is why model choice is the biggest lever on your bill.
Which model for which job
Haiku suits classification, extraction and routing at high volume. Sonnet is the usual starting point for writing, summarizing, customer replies and most analysis. Opus and Fable are worth their price only on hard reasoning or agentic work where a cheaper model fails.
Watch the Haiku threshold
Haiku 5.5 has a price step: prompts above 100,000 tokens are billed at five times the base rate. Sonnet, Opus and Fable include the full 1 million token window at one flat rate. Read more in long-context price cliffs.
How the four tiers relate
The ladder is easy to remember. Opus 5.5 costs twice as much as Sonnet 5.5. Fable 5.1 costs five times as much as Sonnet. Sonnet costs twenty times as much as Haiku 5.5 on input. On every Claude model, output tokens cost five times the input price, so a long answer weighs more on the bill than a long question.
Three realistic monthly budgets
A small app. 200 requests a day, 2,000 input and 500 output tokens each, over 30 days. That is 12 million input and 3 million output tokens a month. Haiku 5.5 costs about $2.70, Sonnet 5.5 about $54, Opus 5.5 about $108 and Fable 5.1 about $270.
A support bot. 3,000 conversations a day, 3,000 input and 300 output tokens each. That is 270 million input and 27 million output tokens a month. Haiku costs about $40, while Sonnet costs about $810. If a cheap model resolves most tickets, the saving is large.
A bulk content job. 100,000 documents, 1,000 input and 300 output tokens each, one time. Sonnet at standard rates costs about $500, or $250 through the batch API. Haiku costs about $25, or about $12.50 in batch.
How Claude pricing has moved
Prices have mostly gone down. Haiku 5.5 launched at $0.10 / $0.50 for prompts up to 100,000 tokens, against $1 / $5 for Haiku 4.5. Sonnet 5.5 kept the same $2 / $10 as Sonnet 5, and Opus 5.5 was priced about 20% below Opus 5. Cache reads on Sonnet 5.5 were also reduced. Because prices change quickly, treat any number in a blog post, including this one, as a snapshot and check the official pricing page before you commit.
API billing vs subscriptions
The API is billed per token and is separate from the consumer and team subscriptions. Some plans have been reported to include monthly API credits, so check your plan before assuming you will pay the full list price. For production workloads, estimate on list prices and treat any credits as a bonus.
Frequently asked questions
Which Claude model is cheapest?
Claude Haiku 5.5, at $0.10 input and $0.50 output per million tokens for prompts up to 100,000 tokens.
Do I pay extra for a long prompt?
On Sonnet 5.5, Opus 5.5 and Fable 5.1, the full 1 million token window is billed at one rate. Haiku 5.5 steps up to $0.50 / $2.50 above 100,000 tokens.
How can I lower my Claude bill?
Use the smallest model that does the job, cache repeated context and batch non-urgent work. See 7 ways to cut your LLM bill.
Is Opus worth it?
Only when quality gains are worth double the Sonnet price. Run a sample of your real tasks on both and compare the success rate, not just the price.
Estimate your own bill
Prices only matter against your real usage. Put your numbers into the free LLM cost calculator to see the monthly cost of every Claude model next to GPT, Gemini, Grok and others.
Prices are list prices checked on November 4, 2026 and change often. Always confirm on Anthropic's pricing page.
Track spend by team and project: the AI Spend Tracker spreadsheet is launching soon.