Claude Opus 5.5 API pricing
Claude Opus 5.5 costs $4 per million input tokens and $20 per million output tokens on Anthropic's API. Input read back from the prompt cache costs $0.20 per million. A typical support chat reply (1,500 tokens in, 300 out) costs about $0.012.
| Input | $4 per 1M tokens |
|---|---|
| Output | $20 per 1M tokens |
| Cached input | $0.20 per 1M tokens |
| Cache writes | $5 per 1M tokens |
| Context window | 1M tokens |
| Reads | text, images, files |
| Released | September 22, 2026 |
Claude Opus 5.5. Counts about 30% more tokens for the same text than Claude models before 4.7, so the same request uses more tokens (Anthropic's pricing page).
What requests cost on Claude Opus 5.5
The same four requests on every page, so you can compare models directly.
| Request | Tokens in / out | Per request | Per 10,000 requests |
|---|---|---|---|
| A support chat reply | 1,500 / 300 | $0.012 | $120 |
| Summarizing a 10-page document | 6,000 / 500 | $0.034 | $340 |
| Sorting a support ticket into a category | 400 / 5 | $0.0017 | $17.00 |
| A question about a 120-page document | 60,000 / 800 | $0.256 | $2,560 |
Anthropic's cheapest model here, Claude Haiku 5.5, costs 98% less for the same chat reply. Whether it is good enough depends on the request: simple ones (sorting, short summaries, extraction) often are, hard reasoning often isn't.
Other Anthropic models
| Model | Input | Output | A chat reply | Next to Claude Opus 5.5 |
|---|---|---|---|---|
| Claude Haiku 5.5 | $0.10 | $0.50 | $0.00030 | 98% less |
| Claude Haiku 4.5 | $1 | $5 | $0.0030 | 75% less |
| Claude Sonnet 5.5 | $2 | $10 | $0.0060 | 50% less |
| Claude Sonnet 5 | $2 | $10 | $0.0060 | 50% less |
| Claude Sonnet 4.6 | $3 | $15 | $0.0090 | 25% less |
| Claude Sonnet 4.5 | $3 | $15 | $0.0090 | 25% less |
| Claude Opus 5 | $5 | $25 | $0.015 | 25% more |
| Claude Opus 4.8 | $5 | $25 | $0.015 | 25% more |
| Claude Opus 4.7 | $5 | $25 | $0.015 | 25% more |
| Claude Opus 4.6 | $5 | $25 | $0.015 | 25% more |
| Claude Opus 4.5 | $5 | $25 | $0.015 | 25% more |
| Claude Fable 5.1 | $10 | $50 | $0.03 | 2.5x the price |
| Claude Fable 5 | $10 | $50 | $0.03 | 2.5x the price |
Similar price from other companies
| Model | Input | Output | A chat reply | Next to Claude Opus 5.5 |
|---|---|---|---|---|
| GPT-5.6 Sol | $4 | $20 | $0.012 | same price |
| GPT-5.5 | $5 | $30 | $0.017 | 38% more |
| GPT-5.4 | $2.50 | $15 | $0.0083 | 31% less |
| GPT-5.2 | $1.75 | $14 | $0.0068 | 43% less |
| GPT-4o | $2.50 | $10 | $0.0067 | 44% less |
Head-to-head
Paying Claude Opus 5.5 prices for every request?
Most apps send simple requests to the same strong model as the hard ones. PromptUnit checks each request your app sends and uses a cheaper model only when it is just as good for that request. You pay 20% of what it saves, and nothing if it saves nothing.
Checked October 14, 2026. Prices are read every day from Anthropic's pricing page. Context window, inputs and release date come from OpenRouter's public model list.
These are the standard list prices on each provider's own API. Batch, flex and priority tiers, cloud platforms such as Amazon Bedrock or Google Vertex AI, and regional endpoints are priced differently. Request costs above count input and output tokens only, not cache, search or image fees.