The development

What happened

Anthropic released Claude Opus 5.5 on 22 September for complex coding, computer use and knowledge work. It reports typical task costs 40% below Opus 5 at default settings, alongside faster output and stronger performance. That task estimate combines efficiency with pricing changes and should be distinguished from the token-rate reductions.

Key facts

  • Standard input and output rates are $4 and $20 per million tokens, respectively, 20% below Opus 5; cache reads cost $0.20 per million, 60% less.

  • Anthropic reports output generation over 30% faster and improved behavioral audit results. These are company findings, not guarantees for every workload.

  • Availability includes Claude Platform, AWS, Google Cloud and Azure under API identifier claude-opus-5-5.

  • The model offers zero-data-retention availability, but thinking mode can no longer be switched off. Cybersecurity and biology safeguards can restrict or reroute requests.

The context

Why it matters

Lower rates can make demanding agent workflows more affordable, but migration also changes thinking-mode and safeguard behavior. Teams should test those integration constraints alongside task quality and cost; better audit scores do not eliminate the need for permissions and containment.