claudecoworkexpert.com
Back to news
September 22, 2026

Claude Opus 5.5: complex knowledge work with lower run costs

Anthropic launched Opus 5.5 for complex coding and knowledge work. The company says a typical task costs around 40% less than with Opus 5 at standard settings.

Anthropic introduced Claude Opus 5.5 on September 22, 2026 as the first model in its Claude 5.5 family. It is aimed at demanding work that requires careful judgement, longer reasoning and multiple steps.

Anthropic reports more than 30% faster output and around 40% lower execution costs on typical workloads at standard settings compared with Opus 5. These are average results from the provider's testing; an individual task may behave differently.

The listed API prices are $4 per million input tokens and $20 per million output tokens, with cache reads at $0.20 per million tokens. Opus 5.5 also includes safeguards designed for more capable models and was evaluated by external reviewers before release.

For teams: try it on a complex analysis or document task. Define what a good result looks like before starting, then compare completeness, correctness and the amount of human correction needed.

Key takeaways

  • Built for complex coding and knowledge work
  • Anthropic reports around 40% lower typical run costs than Opus 5 at standard settings
  • More than 30% faster output according to Anthropic
  • Use a representative task and measure accuracy as well as cost and speed

This site is an initiative of erikvanderveen.nl