Claude Opus 5.5

Anthropic's new top Opus model, performing at the level of Fable 5.1 for about 40% less per task than Opus 5, on all paid Claude plans and the API.

Always thinks before it replies, so Anthropic's guidance is to define what "done" looks like and when to check in, and to drop instructions like "think carefully"
API price is $4 per million input tokens and $20 per million output tokens, 20% below Opus 5, with cached reads at $0.20 per million (60% lower) and a 1M-token context window
Generates output about 30% faster than Opus 5, and Fast mode is available at $8 input and $40 output per million tokens
Five-hour usage limits went up across Pro, Max, Team, and seat-based Enterprise plans, and subscribers got a one-time limit reset to use whenever they want
In Claude Code it is now the default Opus model, 5-hour session limits rose 20%, the lower price stretches limits about 25% further, and Pro and Team Standard plans switched their default from Sonnet to Opus
Available in the Claude apps, on the Claude API as claude-opus-5-5, and on AWS, Google Cloud, and Microsoft Azure

Long, multi-step work you hand off whole, like agentic coding, computer use, and research-heavy knowledge work.

You were running big tasks on Fable 5.1 or Opus 5 and want the same quality faster and cheaper.

Pro, Max, Team, and Enterprise users who hit usage limits, and developers paying per token on the API.

Join the newsletter to get ahead with AI.

I share strategies, new modules, free guides, and tools to get you better at using AI in your work.