Claude Opus 5

Claude Opus 5 is the current Opus-tier model and the one Anthropic’s own documentation tells you to reach for first: “start with Claude Opus 5 for complex agentic coding and enterprise work.” It sits below Claude Fable 5 in raw capability and above Claude Sonnet 5 in price.

What it does

Text and image input, text output, vision, multilingual, and tool use — the same baseline every current Claude model shares. What separates Opus 5 in practice is sustained agentic work: long multi-step tasks with tool calls, file edits and verification loops, where a weaker model drifts or declares victory early.

It has the most recent reliable knowledge cutoff of any current model: May 2026, against January 2026 for both Fable 5 and Sonnet 5.

What it does not do

• No thinking.type: "enabled". Extended thinking as a request parameter is not supported. Opus 5 uses adaptive thinking instead, and decides how much to think for itself. Code written for Claude Sonnet 4.5 or Opus 4.5 that sets thinking explicitly needs updating.
• No temperature, top_p or top_k. Those parameters are deprecated from Claude Opus 4.7 onward and return a 400 error if you set them to a non-default value. Steer with prompting instead.

Specifications

Property	Value
Claude API ID	claude-opus-5
Amazon Bedrock ID	anthropic.claude-opus-5
Google Cloud ID	claude-opus-5
Context window	1M tokens
Max output	128K tokens (300K on the Batch API with the output-300k-2026-03-24 beta header)
Adaptive thinking	Yes
Extended thinking parameter	No
Reliable knowledge cutoff	May 2026
Training data cutoff	May 2026

The ID is a pinned snapshot, not an evergreen alias. From the Claude 4.6 generation onward, dateless IDs still point at one fixed release — see model IDs and versioning.

Pricing

Rate	Price per million tokens
Input	$5
Output	$25
5-minute cache write	$6.25
1-hour cache write	$10
Cache read	$0.50
Batch input / output	$2.50 / $12.50

Fast mode is available on Opus 5 at premium pricing — $10 input and $50 output per million tokens — on the first-party Claude API only.

Two multipliers stack on top of all of the above: inference_geo: "us" applies 1.1x, and regional or multi-region endpoints on Bedrock and Google Cloud carry a 10% premium over global.

Availability

Generally available on the Claude API, Amazon Bedrock, Claude Platform on AWS, Google Cloud and Microsoft Foundry, and used across the Claude apps and Claude Code on paid plans.

On the Claude API and Claude Code, the effort parameter defaults to high for Opus 5. Set it explicitly if you want a different level.

Lifecycle

Active, with a tentative retirement date not sooner than 24 July 2027 — one year after release, which matches the pattern across the current generation. Anthropic commits to at least 60 days’ notice before retiring a publicly released model.