Stories
Anthropic Unveils Claude Opus 5.5: Faster, Cheaper, and Safer AI Model
- •Anthropic launches Claude Opus 5.5, its most powerful AI model, offering stronger performance at significantly lower prices.
- •Opus 5.5 performs on par with Claude Fable 5.1 but costs 40% less for typical workloads, and is faster and writes more clearly.
- •The new model is 85% less likely to bypass prescribed boundaries than Opus 5 or Mythos 5.1, demonstrating improved safety.
- •Early testers reported significant performance jumps, with one writing 680,000 lines of code in less than a day.
- •Opus 5.5 excels in coding, knowledge work, and financial analysis, identifying an error in Walleye Capital's evaluation instructions.
Key Questions Answered
Claude Opus 5.5 requires less compute to serve than its predecessor, resulting in typical workload costs dropping by 40% at default settings, with input tokens priced at $4 and output tokens at $20 per 1M tokens.
Opus 5.5 features a pre-execution screening classifier, an auditable open-source sandbox, vulnerability code reviews, and strong prompt-injection defenses. It also includes anti-distillation safeguards preventing API users from editing prior content to extract reasoning, applying to accounts created on or after August 31, 2026.
Claude Opus 5.5 excels at knowledge work like research, financial analysis, and business modeling, with 16 out of 18 internal research reports clearing quality bars. For coding, an early tester used it to audit and fix a 200,000-line codebase in under three hours.
