Claude Opus 5.5: Flagship Power at Mid-Tier Cost

Three weeks after Fable 5.1, a cheaper Opus model matches it on most work.

Anthropic releases Claude Opus 5.5, which matches Claude Fable 5.1 on most tasks at 40% lower running cost than Opus 5, posts large gains in agentic coding and computer use, and records Anthropic's best alignment-audit results to date.

Key Numbers

Terminal-Bench 4.0

66.4% (Opus 5: 52.3%)

API price

$4 / $20 per M tokens

Verified Facts

• Anthropic said Opus 5.5 "performs at the level of Claude Fable 5.1 on most work and costs 40% less to run than Opus 5," the model it replaced after about two months.
• Published scores against Opus 5: Terminal-Bench 4.0 agentic coding 66.4% vs 52.3%, OSWorld 2.0 computer use 81.8% vs 74.0%, and AutomationBench business workflows 40.0% vs 26.9%. It also generated output more than 30% faster.
• API prices fell to $4 per million input tokens and $20 per million output tokens, from $5 and $25.
• Anthropic reported the best scores of any model to date on its automated behavioral audit, with Opus 5.5 much less likely to take hard-to-reverse actions and more resistant to prompt injection than Opus 5. It was the first model in the Claude 5.5 family, with Sonnet 5.5 and Haiku 5.5 to follow.

Frequently Asked Questions

Sources & Further Reading

• Introducing Claude Opus 5.5 (Anthropic, Sept 22, 2026)
• Anthropic releases Opus 5.5 with lower prices and Fable-level performance (TechCrunch, Sept 22, 2026)

Cite This Page