Claude Opus 5.5 is Anthropic's most capable model in the Opus line, and the one Anthropic recommends starting with for most serious work. It was built for long-running agentic and knowledge work - the dense contract, the gnarly plan, the research thread that runs for an hour.
Released September 22, 2026 by Anthropic. Opus 5.5 has a 1M-token context window and adaptive thinking that is always on - the model decides how much to think before it answers. In Anthropic's lineup it sits between Sonnet 5 and Fable 5.1, at $4 per million input tokens and $20 per million output tokens.
On the All-inclusive plan an Opus 5.5 reply costs 5 credits - the premium tier, the same as Opus 4.6.
Give Opus 5.5 to your strategist agent: goal planning, long-document drafting, contract review, and anything that has to hold a lot in mind at once. For the very hardest problems, Fable 5.1 goes further.