Anthropic Launches Claude Opus 5.5 AI Model with Enhanced Performance and Cost Efficiency

Anthropic has introduced Claude Opus 5.5, its newest flagship AI model designed for coding and knowledge work, boasting lower operational costs and enhanced performance across key benchmarks. This model is part of the Claude 5.5 family and is reported to perform comparably to its predecessor, Claude Fable 5.1, while being 40% cheaper to operate.

The pricing for Claude Opus 5.5 is set at $4 per million input tokens and $20 per million output tokens, which is 20% lower than the previous model, Opus 5. Additionally, cached reads have decreased by 60%, now costing $0.20 per million tokens, and output generation has become over 30% faster.

Benchmark results indicate that Opus 5.5 has made substantial advancements in agentic coding, achieving a score of 66.4% on Terminal Bench 4.0, outperforming both Opus 5 and GPT 6 Astra. The model also excelled in knowledge work, scoring 1,846 Elo on GDPval AA 2.1, surpassing its predecessors.

One early user reported completing a 680,000-line code migration in under a day, while another managed to audit and repair a 200,000-line codebase in less than three hours. In addition to performance improvements, Anthropic has implemented stronger cybersecurity safeguards in Opus 5.5, with a significant reduction in attempts to cross containment boundaries.

Claude Opus 5.5 is now available on various platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure, with additional models in the Claude 5.5 series expected to be released soon.