ArticlesPricing

Claude API Pricing (2026): Per-Million-Token Rates for Every Model

Every current Claude API price per million tokens: input, output, cache writes, cache reads and Batch API rates, plus the subscription break-even and the tokenizer change that raises real bills.

September 20268 min read
claude api pricinganthropic api pricingclaude token costclaude token pricingclaude cost per million tokensclaude batch api pricingclaude prompt caching cost

TL;DRClaude API billing is per million tokens, charged separately for input and output. As of 7 September 2026 the current lineup runs Claude Opus 5 at $5 in and $25 out, Claude Sonnet 5 at $2 in and $10 out, and Claude Haiku 4.5 at $1 in and $5 out. Prompt caching cuts repeat input to 0.1x the base rate, the Batch API halves both directions, and the full 1M-token context window is charged at standard rates. The two things most pricing pages miss: Sonnet 5's $2/$10 was introductory pricing that became permanent, and models from Claude 4.7 onward use a tokenizer that produces roughly 30% more tokens for the same text, so a rate cut can still raise your bill. Sentra, the company brain, is the memory layer that attacks the input side of this equation by sending agents a few governed facts instead of whole documents, and the rates below are laid out first so you can price your own workload.

Claude API pricing per million tokens

All figures are USD per million tokens (written MTok) and come from Anthropic's published pricing, cross-checked against claude.com/pricing on 7 September 2026. Input and output are billed at different rates, and output is always the expensive half.

ModelInput5m cache write1h cache writeCache readOutput
Claude Fable 5.1$10$12.50$20$0.25$50
Claude Fable 5$10$12.50$20$1$50
Claude Opus 5$5$6.25$10$0.50$25
Claude Opus 4.8$5$6.25$10$0.50$25
Claude Opus 4.5$5$6.25$10$0.50$25
Claude Sonnet 5$2$2.50$4$0.20$10
Claude Sonnet 4.6$3$3.75$6$0.30$15
Claude Haiku 4.5$1$1.25$2$0.10$5

Cache reads are 0.1x the base input price on every model except Fable 5.1, where they are 0.025x. Retired models still reachable through Amazon Bedrock and Google Cloud carry their old rates: Claude Opus 4.1 and Opus 4 remain $15 in and $75 out, which is three times the current Opus price for a worse model, so check which model id your code actually pins.

Batch API pricing: half price for asynchronous work

The Batch API applies a flat 50% discount to both input and output for requests you do not need answered immediately. Anything on a schedule rather than in a user's face belongs here: nightly enrichment, backfills, evaluation runs, bulk classification.

ModelBatch inputBatch outputStandard inputStandard output
Claude Opus 5$2.50$12.50$5$25
Claude Sonnet 5$1$5$2$10
Claude Haiku 4.5$0.50$2.50$1$5

What prompt caching actually saves

Prompt caching charges a premium to write context into the cache and then a steep discount to read it back. A 5-minute cache write costs 1.25x the base input rate, a 1-hour write costs 2x, and a read costs 0.1x. That arithmetic gives a clean break-even: the 5-minute cache pays for itself after a single read, and the 1-hour cache pays for itself after two.

So caching is close to free money for an agent that re-reads the same system prompt and the same repository context across a session, and it is a pure loss for one-shot calls that never read the cache back. The multipliers also stack with the Batch discount and with data-residency pricing.

The tokenizer change that makes a rate cut into a price rise

This is the single most under-reported line in Claude API pricing. Models from Claude 4.7 onward use a newer tokenizer that produces roughly 30% more tokens for the same text than the tokenizer used by Sonnet 4.6 and earlier. Rates are charged per token, so identical prompts on a newer model bill for more tokens.

Work an example. Sonnet 4.6 costs $3 per million input tokens; Sonnet 5 costs $2, a 33% cut on paper. Feed both the same 100,000-character prompt and the older model might count 25,000 tokens at $3, which is $0.075, while the newer one counts about 32,500 tokens at $2, which is $0.065. Still cheaper, but 13% cheaper rather than 33%. Move from Sonnet 4.6 to Opus 5 and the same 30% inflation lands on top of a rate increase. Always price a migration on measured token counts from the token counting endpoint, never on the headline rate.

Where the API bill really comes from

Teams reach for a cheaper model when the invoice grows, and that is usually the wrong lever. Two things dominate a real agent bill, and neither is the per-token rate.

  • Output is 5x input on every current model. Opus 5 costs $5 in and $25 out. A verbose agent that narrates its reasoning burns the expensive side of the meter, so response-length discipline beats model downgrades.
  • Input volume compounds with conversation length. Every turn resends the whole history, so token spend on a long agent session grows quadratically rather than linearly. This is the mechanism behind most surprise bills.
  • Retrieval pads the prompt. A RAG pipeline that pastes ten full documents into context to answer one question pays for all ten, every turn, whether or not they were relevant.
  • Retired model ids quietly cost triple. Opus 4.1 at $15 in and $75 out is still callable and still billed at the old rate.

The third item is the one a memory layer fixes. Sentra compiles the answer to a question into a handful of governed, current facts with their provenance, so an agent receives the two sentences that settle the matter instead of the six documents that mention it. Cutting input tokens at the source compounds with prompt caching rather than competing with it, and it removes the stale answers a vector search returns when an old document still matches the query.

API rates or a subscription: the break-even

Claude Code ships inside the paid Claude subscriptions, so the honest question for most developers is which billing model is cheaper at their volume. Take a Pro plan at $20 a month and assume a normal coding mix of roughly 80% input and 20% output tokens by count.

ModelBlended cost per 1M tokensTokens equal to $20/moRoughly per working day
Claude Haiku 4.5$1.8011.1M505k
Claude Sonnet 5$3.605.6M253k
Claude Opus 5$9.002.2M101k

Read that as a threshold, not a recommendation. If your Opus 5 usage runs past roughly 100,000 tokens on a working day, the $20 subscription is already the cheaper route; below it, API billing wins and you only pay for what you use. Add prompt caching and the API side stretches further: assume 70% of input tokens become cache reads and Sonnet 5's blended rate falls to about $2.59 per million, pushing the break-even out to roughly 7.7M tokens a month. Both figures assume 22 working days and no Batch discount.

Costs that sit outside the token meter

  • Web search runs $10 per 1,000 searches on top of the tokens the results consume.
  • Web fetch adds no charge beyond the tokens of whatever it pulls into context.
  • Code execution gives every organization 1,550 free container-hours a month, then bills $0.05 per hour per container.
  • Claude Managed Agents add $0.08 per session-hour, metered only while a session is actually running.
  • Pinning inference to the US with inference_geo applies a 1.1x multiplier to every token category on Claude 4.6 and later.
  • Regional and multi-region endpoints on Bedrock and Google Cloud carry a 10% premium over global endpoints.
  • Tool definitions are input tokens. Declaring the computer-use toolset alone adds about 4,500 tokens to every request that carries it.

Frequently Asked Questions

How much does the Claude API cost per million tokens?
It depends on the model and the direction. As of September 2026, Claude Opus 5 is $5 per million input tokens and $25 per million output tokens, Claude Sonnet 5 is $2 and $10, and Claude Haiku 4.5 is $1 and $5. Output is priced at 5x input across the current lineup. Figures are from Anthropic's pricing documentation.
What does one token cost in Claude?
Divide the per-million rate by a million. A single Sonnet 5 input token costs $0.000002 and an output token costs $0.00001. For practical estimates, a token is roughly four characters or 0.75 words of English, so 1,000 words of input on Sonnet 5 costs about $0.0027.
How much do 1 million tokens cost on Claude?
One million input tokens costs $1 on Haiku 4.5, $2 on Sonnet 5 and $5 on Opus 5. One million output tokens costs $5, $10 and $25 respectively. A million tokens of mixed traffic at a typical 80/20 input-output split blends to roughly $1.80 on Haiku 4.5, $3.60 on Sonnet 5 and $9.00 on Opus 5.
Is Claude Sonnet 5 still $2 per million input tokens?
Yes. The $2 input and $10 output pricing was announced as introductory pricing through 31 August 2026, and Anthropic has confirmed it is now the standard price. The increase to $3 and $15 that had been scheduled for 1 September 2026 will not happen.
Does the 1M-token context window cost extra?
No. On Claude 4.6 and later, the full 1M-token context window is billed at standard per-token rates, so a 900,000-token request costs the same per token as a 9,000-token one. Prompt caching and Batch discounts also apply at standard rates across the whole window.
How much does prompt caching save?
A cache read costs 0.1x the base input rate, against a 1.25x premium for a 5-minute cache write and 2x for a 1-hour write. The 5-minute cache therefore pays for itself after one read and the 1-hour cache after two. For an agent that re-reads the same context all session, caching removes most of the input bill; for one-shot calls that never read the cache back, it is a net loss.
Is the Claude API cheaper than a Claude subscription?
Below roughly 100,000 Opus 5 tokens a working day, or 253,000 Sonnet 5 tokens, API billing costs less than a $20 Pro plan and you pay only for actual use. Past that threshold the subscription is cheaper. Prompt caching moves the line substantially in the API's favour.
Why did my bill go up after switching to a newer Claude model?
Almost certainly the tokenizer. Claude 4.7 and later models use a tokenizer that produces about 30% more tokens for the same text than earlier models. Because billing is per token, a lower headline rate can still mean a higher invoice. Measure with the token counting endpoint before and after any model migration.
How do you reduce Claude API costs without downgrading the model?
Cap output length, since output costs 5x input. Turn on prompt caching for anything re-read within a session. Move non-interactive work to the Batch API for a flat 50% saving. Then cut input volume at the source: a memory layer like Sentra sends an agent the few current facts that answer a question, with provenance, instead of the documents that merely mention it, which shrinks the prompt on every single turn.

Sentralize your company.

Remember what matters.

Resources
Articles
Preferences

Subprocessors include Amazon Web Services, GitHub, Slack, Google Cloud Platform, and OpenAI.

© 2026 Dynamis Labs Inc. All rights reserved.