Models·Anthropic

Claude Opus 5.5: Review and pricing

anthropic/claude-opus-5.5

Claude Opus 5.5 is Anthropic's reasoning model for complex coding and long, multi-step work.

$4Input, per million tokens
$20Output, per million tokens
1MContext window
128KLongest answer

1M tokens of context — about 750,000 words, on a scale that ends at 2M.

What it can do

  • Reasoning
  • Vision
  • Tool use
  • Takes text
  • Takes image
  • Takes file
  • Gives text

Released 22 Sept 2026 · checked 3 Oct 2026.

About it

Claude Opus 5.5 is Anthropic's flagship model, released September 22, 2026. It handles text, images, and files, supports reasoning and tools, and can read up to 1,000,000 tokens in one context. Input costs $4 per million tokens and output costs $20 per million tokens.

Review

In a line

Claude Opus 5.5 is Anthropic's model for demanding reasoning, coding, and long sequences of tasks. It accepts text, images, and files, and returns text.

What it's good at

The model is suited to multi-step reasoning, coding work, and tasks that continue over a long time. Its stated strengths include making changes across large codebases.

It supports tool use and can work with images, so software built around it can provide actions or visual material alongside text. See vision language models for models with image understanding.

Price in plain terms

Input costs $4 per million tokens, while generated output costs $20 per million tokens. A token is a small piece of text used for billing.

At the listed rough rates, reading a 500-word page costs about 0.26 cents. A chat turn with about 1,000 tokens in and 500 out costs about $0.01; 1,000 such turns cost about $14.00.

Context and limits

Claude Opus 5.5 can take up to 1,000,000 tokens in one working context. That is the material it can consider together, such as text, images, or files.

One answer can contain up to 128,000 tokens. Its listed output is text, so the facts do not describe audio or video output.

Ways to use it

People can access it through software or chat tools that offer Claude Opus 5.5. Developers can identify the model with the ID anthropic/claude-opus-5.5.

Because it accepts text, images, and files, an app can send those input types. Tool support allows connected software to give it actions to perform.

Verdict

Choose Claude Opus 5.5 when a task needs careful reasoning, substantial coding changes, or a very large amount of reference material at once. Image and file input may also suit workflows that are not text-only.

Skip it when you need audio or video output, or when the higher output rate does not fit your budget. For other Anthropic options, see Anthropic models and newest models.

Verdict

Claude Opus 5.5 fits people building software or services around difficult, multi-step work. Its million-token context can hold unusually large inputs, and it can process images and files as well as text. The trade-off is cost: generated text is billed at $20 per million tokens, five times the input rate. It is also text-output only in the listed specifications. Consider it when depth and long-running tasks matter more than keeping output spending low.

  • Handles multi-step reasoning and coding tasks
  • Reads up to 1,000,000 tokens at once
  • Accepts text, images, and files
  • Supports tool use
  • Can generate up to 128,000 tokens
  • Output costs $20 per million tokens
  • Listed output is text only
  • Large context may cost more as usage grows
  • Pick it ifComplex coding and reasoning; Long tasks with large reference files; Apps needing image or tool input
  • Skip it ifYou need audio or video output; Your workload produces a lot of generated text; You do not need a large context window

Pricing

Input is $4 per million tokens and output is $20 per million tokens. Roughly, one 500-word page costs 0.26 cents to read, one sample chat turn costs $0.01, and 1,000 sample turns cost $14.00.

Questions

Who makes Claude Opus 5.5?

Anthropic makes Claude Opus 5.5.

What can Claude Opus 5.5 read?

It accepts text, images, and files, with a context window of up to 1,000,000 tokens.

How much does its output cost?

Generated output costs $20 per million tokens. Input costs $4 per million tokens.

How much does a typical chat turn cost?

A rough example with 1,000 input tokens and 500 output tokens costs about $0.01.

Does it support images?

Yes. Vision is listed, and images are one of its accepted input types.

How long can one answer be?

The maximum output is 128,000 tokens.

What does tool support mean here?

Tools are supported, so software connected to the model can provide actions for it to use.

On these boards

More from Anthropic