Calculate Claude API costs by model, article length, monthly volume and supported processing tier.
Catalog rates verified . See verification and correction history.
Base text-token estimate, not an invoice. Excludes reasoning beyond the word estimate, output overshoot, cache writes/storage, tools, regional surcharges and editing. Unsupported tiers use Standard and say so. Check the calculation assumptions and budget for human editing.
One 1,500-word article costs $0.0101 on Claude Haiku 4.5, Anthropic's cheapest tracked model, and $0.1014 on Claude Fable 5, its most expensive. On Claude Haiku 4.5, Batch costs $0.00507 at these assumptions.
This is a price ranking, not a writing-quality recommendation. Check the model-selection guide and real AI-written article examples before deciding what fits your editorial workflow.
One draft, a 250-word brief, 20% prompt overhead. Change any of it in the calculator.
| Model | 500w | 1,000w | 1,500w | 2,000w | 3,000w |
|---|---|---|---|---|---|
| Claude Haiku 4.5 | $0.00364 | $0.00689 | $0.0101 | $0.0134 | $0.0199 |
| Claude Sonnet 5.5 | $0.00728 | $0.0138 | $0.0203 | $0.0268 | $0.0398 |
| Claude Sonnet 5 | $0.00728 | $0.0138 | $0.0203 | $0.0268 | $0.0398 |
| Claude Sonnet 4.6 | $0.0109 | $0.0207 | $0.0304 | $0.0402 | $0.0597 |
| Claude Opus 5.5 | $0.0146 | $0.0276 | $0.0406 | $0.0536 | $0.0796 |
| Claude Opus 5 | $0.0182 | $0.0345 | $0.0507 | $0.0669 | $0.0994 |
| Claude Opus 4.8 | $0.0182 | $0.0345 | $0.0507 | $0.0669 | $0.0994 |
| Claude Fable 5.1 | $0.0364 | $0.0689 | $0.1014 | $0.1339 | $0.1989 |
| Claude Fable 5 | $0.0364 | $0.0689 | $0.1014 | $0.1339 | $0.1989 |
Cheapest first. These are list rates applied to a token model — see the methodology for exactly where that diverges from an invoice.
The 1,000w column is the per-1,000-words figure people quote: $0.00689 on Claude Haiku 4.5 at these settings.
Providers do not price the same way, and the differences change which model wins.
Anthropic prices Claude on a simple ladder: each step up roughly doubles or halves the rate, and input is consistently one fifth of output across the whole range. Cache reads on most Claude models tracked here are 10% of input — Opus 5.5 charges 5% and Fable 5.1 2.5% — so check the model rather than assume a catalog-wide rule. Cache creation costs extra and is excluded from this estimate, so a cache-hit percentage alone is not the full cache budget.
All tracked Claude models offer Standard and Batch, with eligible Batch token charges 50% off. Opus 5.5, Opus 5 and Opus 4.8 also offer Fast mode at 2× standard rates on the first-party API. Fast and Batch cannot be combined; there is no Claude Flex tier in this calculator.
Every current Claude model except Haiku 4.5 carries a 1M-token context window, and Anthropic does not charge a higher rate above a context threshold the way Google and xAI do. An article workflow never approaches the limit, so the rate you see is the rate you pay.
No tracked Claude model currently has an announced rate change. When one is announced, the footnote in the table below states it before it takes effect. Why the announced Sonnet 5 price increase was cancelled.
Every gallery below includes an unedited Claude Opus 5 and Claude Sonnet 5 draft, published blind with a current-rate cost comparison. The model and the price stay hidden until you reveal them, so this page can name who took part without spoiling which draft is which.
USD per million tokens, cheapest output first.
| Model | Input | Cached in | Output | Context |
|---|---|---|---|---|
| Claude Haiku 4.5 | $1.00 | $0.100 | $5.00 | 200K |
| Claude Sonnet 5.5 | $2.00 | $0.200 | $10.00 | 1,000K |
| Claude Sonnet 5 | $2.00 | $0.200 | $10.00 | 1,000K |
| Claude Sonnet 4.6 | $3.00 | $0.300 | $15.00 | 1,000K |
| Claude Opus 5.5 | $4.00 | $0.200 | $20.00 | 1,000K |
| Claude Opus 5 | $5.00 | $0.500 | $25.00 | 1,000K |
| Claude Opus 4.8 | $5.00 | $0.500 | $25.00 | 1,000K |
| Claude Fable 5.1 | $10.00 | $0.250 | $50.00 | 1,000K |
| Claude Fable 5 | $10.00 | $1.000 | $50.00 | 1,000K |
Source: Anthropic's published pricing. Compare against the other providers on the full pricing table.
Answers specific to this provider's rate structure.
Claude Haiku 4.5 is the cheapest tracked Claude model for a full-length draft. It is the only current Claude model with a 200K rather than 1M context window, which is irrelevant for article work — a 1,500-word article and its brief use a few thousand tokens.
Roughly two thirds of the 1,500-word figures on this page, because output scales linearly with article length while the brief does not. Anthropic charges output at exactly five times input on every current model, so a longer article costs proportionally more than a longer brief.
Yes. Every tracked Claude model supports Batch at 50% off eligible token charges. Opus 5.5, Opus 5 and Opus 4.8 also support Fast mode at 2x standard rates. Fast cannot be combined with Batch; no Claude Flex tier is modeled here.
Cache reads are 10% of input on most Claude models tracked here, but 5% on Opus 5.5 and 2.5% on Fable 5.1. Cache creation is an additional charge excluded from this estimate. Check the model’s read and write rates before budgeting repeated style guides or long research prompts.
It depends entirely on which models you compare. Claude’s cheapest tracked model costs more per article than OpenAI’s cheapest, but OpenAI’s most expensive costs several times Claude’s most expensive. Compare like for like on the full pricing table rather than by provider.
Price your workflow here, then check it against the other providers before committing. At article volumes the difference between two reasonable models is usually smaller than a single hour of editing. When the work can run asynchronously, compare Claude's batch rate — not its standard rate — against DeepSeek, where no Batch tier is published. DeepSeek instead offers a separate off-peak discount.