Anthropic released Claude Opus 5.5. It matches Fable 5.1 on most work. Input tokens cost $4 per million and output tokens $20 per million. Cache reads are 60% cheaper. Output is over 30% faster than Opus 5. It follows instructions better and communicates more clearly. It still buries the point in writing and can overrun time on tasks. It launches with Fable-class cybersecurity and biology safeguards. Most cybersecurity tasks are redirected to Opus 4.8. It is available in Claude apps, Claude Code, and major cloud platforms. Usage limits for Pro, Max, and Team subscribers are raised. Some team members now prefer it over Codex for daily work. Others still favor Codex or Fable for specific tasks.
What changed
Opus 5 had personality issues, ran long on simple tasks, and ignored instructions in workflows.
What it unlocks
Using a Fable-level model at roughly 60% lower cost for coding, design, and writing tasks.
- 60% cheaper than Fable token price
- $4 per million input tokens
- $20 per million output tokens
- 30% faster output than Opus 5
What you need to act on it
- access to Claude apps, Claude Code, or developer platform
- Pro, Max, or Team subscription for higher usage limits
Sources