Claude Opus 4.6 — pricing, context window and what it's actually good for

Claude Opus 4.6 is an Active Claude model on the Claude API. 1M-token context, 128k max output. Training data cutoff August 2025. Uses the tokenizer used by Claude Sonnet 4.6 and earlier, so per-token costs are directly comparable with models of that generation. Tentative retirement date: not sooner than 5 February 2027 (Anthropic model deprecations, checked 2026-08-07 11:51 UTC).

Context window: 1,000,000 tokens
Max output: 128,000 tokens
Released: 5 Feb 2026
Knowledge cutoff: May 2025
API identifier: claude-opus-4-6

Pricing · per 1M tokens

Input$5.00

Output$25.00

Cached input$0.50

Cached input is 10× cheaper on this model — it pays off when you send the same context again and again.

US data residency · 1.1× list

Input$5.50

Output$27.50

Cached input$0.55

Pinning inference to the US multiplies every token category — input, output, cache writes and cache reads — by 1.1. Global routing is the default and bills at the rates above. Anthropic pricing docs add that the partner-operated platforms — Amazon Bedrock and Google Cloud — have independent regional pricing, so check theirs rather than assuming these figures carry over.

Checked against Anthropic pricing docs on 7 Aug 2026, 11:51 UTC

Best for

Code Generation Reasoning & Research

Sources: Anthropic pricing docs · Anthropic models overview · Anthropic model deprecations

Not sure Claude Opus 4.6 fits your workload?

We'll size the right model for your use case and give you a real cost estimate. One free call, no slides.

Talk to us