Claude Opus 5.5: Anthropic introduces a new model

Anthropic announces Opus 5.5, the first model in its Claude 5.5 family.

Main report

Anthropic announced Opus 5.5 on September 22. Standard API pricing is $4 per million input tokens and $20 for output; Fast mode has separate pricing.

Assess an AI assistant at task level. A faster or cheaper model does not necessarily produce better outcomes for every user. Input quality, permissions and checking all affect the result.

Analysis

TBN Grow analysis: in long tasks, users need to know what changed, why and what verifies correctness. A website edit should be checked alongside related functionality, not judged by a polished demonstration alone.

What to watch

Keep a backup before tasks that alter data or code. Grant only necessary permissions, use checkable stages and ask for unresolved work to be reported. Compare quality on your own tasks before migrating an entire workflow.