TechCompare LogoTechCompare

Claude Opus 5.5 API pricing and cost calculator

An Opus request with 100K input and 5K output costs $0.50 uncached or $0.31 with half the input read from cache. Include cache writes when estimating a full session.

Claude Opus 5.5 costs $4 per million input tokens and $20 per million output tokens at standard rates. Cache reads cost $0.20 per million, making reused context much cheaper than sending it as uncached input each time. That can help an agent with a stable prompt prefix, but it doesn't make its new messages or output free. This page separates those billing categories and shows how a request estimate becomes a monthly budget.

By TechCompare · Updated

Input tokens
100,000
per request
Output tokens
5,000
per request
Volume
100 / monthly
Standard API

Calculator

Cost Comparison

Based on 100,000 input tokens (50% cached), 5,000 output tokens, and 100 requests.Prices are fetched live from OpenRouter and may include temporary promotional discounts not accounted for in our article and comparison figures.

Same 100,000 in / 5,000 out (50% cached) scenario, pre-loaded. Adjust tokens, caching, or swap models.
Edit in the full tool

How this is calculated

A 100,000-input, 5,000-output request costs $0.40 for uncached input and $0.10 for output, or $0.50. The starting calculator example reads half the input from cache: 50,000 uncached tokens cost $0.20, 50,000 cache-read tokens cost $0.01, and output costs $0.10. That totals $0.31 per call or $31 for 100 monthly calls. Opus uses these standard rates within its 1M context window, without a separate long-prompt token surcharge. A 300,000-input, 5,000-output request costs $1.30 uncached or $0.73 with half the input cached. Cache creation isn't included in either example. The embedded calculator uses live OpenRouter provider prices, while the prose describes Anthropic's standard rates.

Verdict

Opus's output rate deserves as much attention as its input discount. A session can reuse a large cached prefix and still spend heavily on reasoning, lengthy responses, or repeated tool turns. Evaluate the cost of an accepted result, then compare it with Sonnet or Sol on the same task. Use Opus where its measured improvement offsets the additional spending, and cap avoidable context growth and retries before increasing request volume.

More API Standalones scenarios

MiMo-V2.6-Pro Pricing
Cost calculator for this model
View details ➜
MiMo-V2.6-Flash Pricing
Cost calculator for this model
View details ➜
GLM-5.3-Flash Pricing
Cost calculator for this model
View details ➜

Related guides

Sol vs Opus 5.5
Compare the input, output, and long-prompt price differences.
View guide ➜
Opus 5.5 vs Sonnet 5.5
See why equal cache-read rates don't mean equal total costs.
View guide ➜
Official Opus pricing reference
Anthropic's input, output, cache-write, and cache-read rates.
View guide ➜
Claude pricing rules
Verify batch discounts, long-context billing, and tool charges.
View guide ➜

Frequently asked questions

How much do 10,000 example Opus 5.5 calls cost?
At 100,000 input and 5,000 output tokens per request, the standard token estimate is $5,000 without cache hits or $3,100 with half the input cached. If all input tokens qualify as cache reads, the token estimate is $1,200. Those totals exclude the cache writes that created or refreshed the cache, plus tool fees and any extra calls.
What is the cost of writing an Opus prompt cache?
A five-minute cache write costs $5 per million tokens, while a one-hour write costs $8. Reading from that cache costs $0.20 per million. Treat writes and reads as separate usage categories and estimate how often the prefix will expire or change. A cache used once has different economics from one reused across hundreds of calls.
Does Opus 5.5 charge more for a long prompt?
The standard per-token rates apply within its 1M context window without a separate long-context premium. A bigger prompt still costs more because it contains more tokens. Processing 300,000 uncached input tokens costs $1.20 before output, for example. Reusing cached context can reduce that input spending, but only the tokens billed as cache reads get the discount.
How does the Opus 5.5 batch discount work?
Anthropic's Batch API reduces applicable token prices by 50%. The uncached 100K-input, 5K-output example therefore becomes $0.25, while the half-cached example becomes $0.155 before write costs and tools. Use batch processing for work that can wait for asynchronous results. The calculator's toggle models the provider's discount rather than an interactive OpenRouter discount.
Can caching make Opus cheaper than Sonnet for identical usage?
Their cache-read rates tie, but Opus has higher rates for uncached input, output, and cache writes. Equal token usage therefore doesn't give Opus a price advantage. It can still be the better task-level choice if it saves enough failed attempts or manual work in your evaluations. Measure those savings rather than inferring them from the model's price or name.

Related tools

LLM VRAM Calculator

Calculate the VRAM needed to run or fine-tune any LLM at any quantization.

Use tool ➜

Power Cost Estimator

Estimate annual electricity costs for your PC, Server, or TV.

Use tool ➜

Data Transfer Calculator

Estimate transfer times for files over USB, WiFi, Ethernet, and more.

Use tool ➜

JSON Formatter

Validate, format, and minify JSON data with readable output and error detection.

Use tool ➜