Claude Opus Pricing: Every Opus Model's API Cost

Claude Opus pricing for every model: Opus 5.5, Opus 5 and Opus 4.5 to 4.8 per million tokens, plus caching, batch, fast mode, a worked example and plans.

Josh Pocock7 min read

Opus pricing cover

AI Coding

Opus pricing

Stride AI Academy

Claude Opus pricing on the API starts at $4 per million input tokens and $20 per million output tokens for Claude Opus 5.5, the newest Opus model. Opus 5 and Opus 4.5, 4.6, 4.7 and 4.8 all cost $5 input and $25 output. Prompt caching, the Batch API and fast mode change those rates, and on a Claude plan you get Opus from Pro upwards with no per-token bill.

This guide has the full Opus price table, caching and batch rates, fast mode prices, a worked cost example, and which Claude plans include Opus. Everything here was checked against Anthropic's pricing docs, models overview, model deprecations, rate limits and Claude's pricing page on November 22, 2026. Prices are in US dollars per million tokens (MTok).

Claude Opus API price by model

From the pricing docs on November 22, 2026:

Model Input Output 5-min cache write 1-hour cache write Cache hit
Opus 5.5 $4 $20 $5 $8 $0.20
Opus 5 $5 $25 $6.25 $10 $0.50
Opus 4.8 $5 $25 $6.25 $10 $0.50
Opus 4.7 $5 $25 $6.25 $10 $0.50
Opus 4.6 $5 $25 $6.25 $10 $0.50
Opus 4.5 $5 $25 $6.25 $10 $0.50
Opus 4.1 (retired) $15 $75 $18.75 $30 $1.50
Opus 4 (retired) $15 $75 $18.75 $30 $1.50

So the newest Opus is also the cheapest. Opus 5.5 is 20% cheaper than Opus 5 on input and output, and its cache hits cost 0.05x the input price instead of the standard 0.1x.

Claude Opus 4.8, 4.7, 4.6 and 4.5 pricing

All four cost $5 per million input tokens and $25 per million output tokens, with the same caching and batch rates. A few differences that affect your bill:

  • Tokenizer. The pricing docs say Claude 4.7 and later models use a newer tokenizer that makes about 30% more tokens for the same text. Opus 4.6 and 4.5 use the older one. So the same prompt can cost more on Opus 4.7 or 4.8 than on 4.6 at the same rate.
  • Context. Opus 4.6, 4.7 and 4.8 have a 1M-token context window and 128K max output. Opus 4.5 has 200K and 64K.
  • Shared rate limit. Opus 4.8, 4.7, 4.6 and 4.5 share one combined rate limit bucket. Opus 5.5 and Opus 5 each have their own, per the rate limits docs.

Retired and retiring Opus models

The deprecations page lists Opus 4.1 as retired on August 5, 2026 and Opus 4 as retired on June 15, 2026. The pricing docs note they may still be available on some cloud platforms. Opus 4.5 is active, with a tentative retirement date of not sooner than November 24, 2026. If you're still on it, plan a move.

Opus caching and batch pricing

Prompt caching. A 5-minute cache write costs 1.25x the input price and a 1-hour write costs 2x. Cache reads cost 0.05x on Opus 5.5 ($0.20) and 0.1x on the other Opus models ($0.50). The docs say a 5-minute cache pays off after one read, and a 1-hour cache after two.

Batch API. Batch requests are 50% off input and output:

Model Batch input Batch output
Opus 5.5 $2 $10
Opus 5, 4.8, 4.7, 4.6, 4.5 $2.50 $12.50

Batch and caching discounts stack. For how caching and batch work across all models, see our Claude API pricing guide.

Opus fast mode pricing

Fast mode gives up to 2.5x higher output speed at a premium:

Model Fast mode input Fast mode output
Opus 5.5 $8 $40
Opus 5 and Opus 4.8 $10 $50

The pricing docs say fast mode is in research preview for Opus 5 and deprecated for Opus 4.8. Fast mode pricing covers the full context window, caching multipliers apply on top, and it isn't available with the Batch API. For the speed question in Claude Code, see how to change the model in Claude Code.

Other multipliers

  • US-only inference. Setting inference_geo: "us" costs 1.1x on Claude 4.6 and later models, so it applies to Opus 5.5, 5, 4.8, 4.7 and 4.6.
  • Long context. Opus 4.6 and later include the full 1M context at standard pricing. A 900K-token request is billed at the same per-token rate as a 9K one.

Claude Opus cost: a worked example

Say an agent run sends 2 million input tokens and gets 200,000 output tokens back. With caching, 1.5 million of the input tokens are cache reads and 500,000 are uncached.

On Opus 5.5:

Line Calculation Cost
Uncached input 0.5 MTok × $4 $2.00
Cache reads 1.5 MTok × $0.20 $0.30
Output 0.2 MTok × $20 $4.00
Total $6.30

The same run on Opus 4.8:

Line Calculation Cost
Uncached input 0.5 MTok × $5 $2.50
Cache reads 1.5 MTok × $0.50 $0.75
Output 0.2 MTok × $25 $5.00
Total $8.25

Both models use the newer tokenizer, so the token counts are comparable. Moving from Opus 4.8 to 5.5 cuts this run by about 24%. Without caching, the Opus 5.5 run would cost $8.00 for input alone instead of $2.30, which is why caching matters more than the model choice for long agent sessions.

Run the batch version and the Opus 5.5 totals halve again, as long as the work can wait. To estimate your own usage in Claude Code, try the Claude Code cost calculator.

Which Claude plans include Opus

On a subscription you don't pay per token. From the pricing page on November 22, 2026:

Plan Price Opus?
Free $0 No
Pro $20/month, or $17/month billed yearly Yes
Max $100/month Yes, with 5x Pro usage
Ultra $200/month Yes, with 10x Pro weekly usage
Team From $20/seat/month billed yearly Yes
Enterprise $20/seat/month billed yearly plus usage at API rates Yes

Opus uses your plan's allowance faster than lighter models. For how the limits work, see Claude usage limits, and for the plans themselves, Claude pricing.

In Claude Code, per the model configuration docs, the opus alias points to Opus 5.5 on the Anthropic API, and the default model on Pro, Max, Ultra, Team and Enterprise is Opus 5.5.

Is Opus worth the price?

The models overview says to start with Opus 5.5 for most workloads if you're unsure. It's built for long-running agentic coding and knowledge work, with adaptive thinking always on and a default effort of medium.

Sonnet 5.5 costs half as much at $2 input and $10 output, and it's the faster model. For a side-by-side, read Claude Sonnet vs Opus and our Claude Sonnet pricing breakdown. Fable 5.2, the step above Opus, costs $8 input and $40 output.

Before you switch models to save money, try lowering effort. The choosing-a-model docs say tuning effort is often a better lever than switching models.

Learn to build with Claude Code

The Stride AI Coding course teaches Claude Code and Codex for building client tools, written for non-programmers, with weekly live calls to review your builds. Many of the free guides cover Claude Code setups, and the Claude Code community is where members compare theirs.

FAQ

How much does Claude Opus cost?

On the API, Claude Opus 5.5 costs $4 per million input tokens and $20 per million output tokens on November 22, 2026. Opus 5 and Opus 4.5 to 4.8 cost $5 and $25. On a Claude plan, Opus is included from Pro at $20 a month.

What is Claude Opus 4.6 pricing?

Opus 4.6 costs $5 per million input tokens and $25 per million output tokens. Cache hits cost $0.50 and batch requests are half price at $2.50 and $12.50. It uses the older tokenizer, so the same text makes fewer tokens than on Opus 4.7 or later.

Is Claude Opus 4.5 pricing the same as Opus 4.8?

Yes. Both cost $5 input and $25 output per million tokens, with the same cache and batch rates. Opus 4.5 has a smaller 200K context window and a tentative retirement date of not sooner than November 24, 2026.

Why is Opus 5.5 cheaper than Opus 4.8?

Anthropic priced Opus 5.5 at $4 and $20 against $5 and $25 for Opus 4.8, and gave it a lower cache read multiplier of 0.05x. The docs don't explain the pricing decision, but the result is that the newest Opus costs the least.

Does the Claude Free plan include Opus?

No. Claude's pricing page lists Sonnet and Haiku on Free. Opus starts on Pro at $20 a month, or $17 a month billed yearly.

How much is Opus fast mode?

Fast mode on Opus 5.5 costs $8 per million input tokens and $40 per million output tokens. On Opus 5 and Opus 4.8 it costs $10 and $50. It can't be combined with the Batch API.

Build it with us in Stride AI Academy

Four courses, weekly live calls and a community of people building and selling AI. Or start with 320+ free guides.

Keep reading

Free AI Coding guides