Introducing Claude Opus 4.5

Anthropic released Claude Opus 4.5, their most advanced AI model optimized for coding, autonomous agents, and computer use. The model achieves state-of-the-art performance on software engineering benchmarks while using 48-76% fewer tokens than previous versions. Key improvements include better reasoning, enhanced safety against prompt injection attacks, and new developer features like an effort parameter for controlling compute tradeoffs. Pricing dropped to $5/$25 per million tokens, making frontier capabilities more accessible. The model outperformed all human candidates on Anthropic's internal engineering exam and excels at long-horizon autonomous tasks, multi-step reasoning, and code generation across various enterprise applications.