Anthropic prices every Claude model per million tokens, with separate rates for input, cached input, cache writes and output, and a 50% Batch API discount. The table below is the full current list, checked against Anthropic's own pricing page, followed by what three real workloads cost on each model.
claude-opus-5-5claude-sonnet-5-5claude-haiku-5-5Cache write is the 5-minute cache; the 1-hour cache is 2x input. Cache reads are 0.1x input on most models, 0.05x on Opus 5.5 and Sonnet 5.5, 0.025x on Fable 5.1 and Mythos 5.1. Batch API is 50% off. Haiku 5.5 prompts over 100k tokens bill at $0.50 / $2.50. Source: platform.claude.com →
| Model | Input | Cached input | Cache write | Output | Batch | Notes |
|---|---|---|---|---|---|---|
Claude Fable 5.1claude-fable-5-1 |
$10 | $0.25 | $12.5 | $50 | 50% off | |
Claude Mythos 5.1claude-mythos-5-1 |
$10 | $0.25 | $12.5 | $50 | 50% off | Limited availability. |
Claude Fable 5claude-fable-5 |
$10 | $1 | $12.5 | $50 | 50% off | |
Claude Mythos 5claude-mythos-5 |
$10 | $1 | $12.5 | $50 | 50% off | Limited availability. |
Claude Opus 5.5claude-opus-5-5 |
$4 | $0.2 | $5 | $20 | 50% off | |
Claude Opus 5claude-opus-5 |
$5 | $0.5 | $6.25 | $25 | 50% off | |
Claude Opus 4.8claude-opus-4-8 |
$5 | $0.5 | $6.25 | $25 | 50% off | |
Claude Opus 4.7claude-opus-4-7 |
$5 | $0.5 | $6.25 | $25 | 50% off | |
Claude Opus 4.6claude-opus-4-6 |
$5 | $0.5 | $6.25 | $25 | 50% off | |
Claude Opus 4.5claude-opus-4-5 |
$5 | $0.5 | $6.25 | $25 | 50% off | |
Claude Sonnet 5.5claude-sonnet-5-5 |
$2 | $0.1 | $2.5 | $10 | 50% off | |
Claude Sonnet 5claude-sonnet-5 |
$2 | $0.2 | $2.5 | $10 | 50% off | |
Claude Sonnet 4.6claude-sonnet-4-6 |
$3 | $0.3 | $3.75 | $15 | 50% off | |
Claude Haiku 5.5claude-haiku-5-5 |
$0.1 | $0.01 | $0.125 | $0.5 | 50% off | Prompts over 100k tokens bill at $0.50 input / $2.50 output. |
List prices are hard to feel per million. These are three monthly workloads, priced on the models above at the standard tier, with the batch discount applied only where the workload is asynchronous. Retries, tool-call loops and prompt growth are not in these numbers, which is why estimates and invoices disagree.
| Model | Support chatbot | Tool-using agent | Overnight classification |
|---|---|---|---|
Claude Fable 5.1claude-fable-5-1 |
$2,000 | $2,925 | $5,250 |
Claude Opus 5.5claude-opus-5-5 |
$800 | $1,170 | $2,100 |
Claude Sonnet 5.5claude-sonnet-5-5 |
$400 | $585 | $1,050 |
Claude Haiku 5.5claude-haiku-5-5 |
$20.00 | $29.25 | $52.50 |
Take the tool-using agent above on Claude Sonnet 5.5. Its 180,000,000 input tokens a month cost $360 at the input rate. If 80% of those tokens are a repeated prefix (system prompt, tool definitions, a shared document) served from the cache at $0.1, the input line falls to $86.40, a 76% saving before output tokens. Anthropic is the one provider that charges to populate the cache: a 5-minute cache write bills at 1.25x the input rate (a 1-hour cache at 2x). The write happens once per distinct prefix, so for a prompt read many times the premium is small next to the saving.
Most Claude models read cached input at 0.1x the input price; Opus 5.5 and Sonnet 5.5 read at 0.05x, and Fable 5.1 and Mythos 5.1 at 0.025x. A system prompt, tool definitions or a document that repeats across calls costs a fraction of its list price after the first request.
Prompts over 100,000 tokens bill at $0.50 input and $2.50 output instead of $0.10 and $0.50. Agents that stuff a long conversation into Haiku to save money can cross that line without noticing.
Input, cached input and output are all 50% off for asynchronous jobs that can wait up to 24 hours. Classification, extraction and evaluation runs that do not need an answer now are the obvious candidates.
Opus 4, Opus 4.1 and Sonnet 4 are retired on the Claude API but still available on Amazon Bedrock and Google Cloud at the prices shown. The dated ids (claude-sonnet-4-5-20250929, for example) price the same as the undated row.
Still priced because code still names them. A legacy model is available but superseded; a retired one no longer runs on this provider's API, though several remain on cloud marketplaces at the rate shown.
| Model | Input | Cached input | Output | Notes |
|---|---|---|---|---|
Claude Opus 4.1retiredclaude-opus-4-1 |
$15 | $1.5 | $75 | Retired on the Claude API; still on Bedrock and Google Cloud. |
Claude Opus 4retiredclaude-opus-4 |
$15 | $1.5 | $75 | Retired on the Claude API; still on Google Cloud. |
Claude Sonnet 4.5legacyclaude-sonnet-4-5 |
$3 | $0.3 | $15 | |
Claude Sonnet 4retiredclaude-sonnet-4 |
$3 | $0.3 | $15 | Retired on the Claude API; still on Bedrock and Google Cloud. |
Claude Haiku 4.5legacyclaude-haiku-4-5 |
$1 | $0.1 | $5 | |
Claude Haiku 3.5retiredclaude-3-5-haiku |
$0.8 | $0.08 | $4 | Retired on the Claude API; still on Google Cloud. |
The same three roles on each provider, input and output per million tokens. Tier for tier is the only fair comparison; a budget model on one provider against a flagship on another tells you nothing.
| Provider | Flagship | Workhorse | Budget |
|---|---|---|---|
| Anthropic | $4 / $20 Claude Opus 5.5 |
$2 / $10 Claude Sonnet 5.5 |
$0.1 / $0.5 Claude Haiku 5.5 |
| OpenAI | $5 / $30 GPT-5.5 |
$2.5 / $15 GPT-5.4 |
$0.05 / $0.4 GPT-5 Nano |
| Google Gemini | $2 / $12 Gemini 3.1 Pro Preview |
$0.75 / $3.75 Gemini 3.8 Flash |
$0.3 / $2.5 Gemini 3.5 Flash-Lite |
Current Claude models range from $0.1 input and $0.5 output per million tokens on Claude Haiku 5.5 to $10 input and $50 output on Claude Mythos 5. The mid-range Claude Sonnet 5.5 is $2 input and $10 output. Cached input is 0.1x or less on every model and the Batch API is 50% off.
Claude Pro and Max are consumer subscriptions for the Claude app, billed monthly per person. The Claude API is billed per token to developers who call the models from their own software. This page covers the API. A heavy API user and a Claude Pro subscriber can spend very different amounts for the same model.
On Claude Sonnet 5.5, a prompt that is 80% cached input costs 76% less in input tokens than the same prompt uncached ($360 against $86.40 a month on the worked agent example below). The write premium is 1.25x input on the first request for each distinct prefix.
It depends which models you line up. At the budget end Claude Haiku 5.5 ($0.1 / $0.5) is priced alongside GPT-5 Nano and Gemini Flash-Lite. At the top end Opus 5.5 is cheaper per token than GPT-5.5 and more expensive than Gemini 3.1 Pro. The cross-provider table on this page puts the equivalent tiers side by side; the full catalogue is on the LLM API pricing comparison.
Extended thinking tokens bill as output tokens at the model's output rate. They do not appear in the visible answer but they are in the usage block of every response, which is where any cost calculation has to start.
These are the rates. Agent Ping applies them to every real run your agents make, by agent, customer and feature, and pages you when one of them spikes. It prices Claude calls from this same table.
Free plan, no card. Prices on this page verified 17 October 2026.