Anthropic launches Claude Opus 5 AI model, priced at half of Fable 5

Claude Opus 5 is designed for enterprise coding, scientific research and knowledge work, and outperforms its predecessor, Opus 4.8.Anthropic on Friday announced the launch of Claude Opus 5, a new artificial intelligence (AI) model that it said delivers performance close to its flagship Claude Fable 5 model at half the price, while becoming the default model for Claude Max subscribers and the most capable model available on Claude Pro.The company said Claude Opus 5 is designed for enterprise coding, scientific research and knowledge work, and outperforms its predecessor, Opus 4.8, while retaining the same pricing of $5 per million input tokens and $25 per million output tokens.According to the company’s statement, Opus 5 achieves state-of-the-art performance on coding and knowledge work benchmarks including Frontier-Bench and GDPval-AA, while remaining behind Mythos 5 on cybersecurity and biology-related tasks.The company said the model is built to help enterprises deploy AI more efficiently by improving verification and iterative reasoning, and includes safeguards designed for cybersecurity and biology applications.

Anthropic added that it has not specifically trained Opus 5 on cybersecurity tasks, although its general capabilities have improved its ability to identify software vulnerabilities.Alongside the model launch, Anthropic introduced mid-conversation tool changes on the Claude Platform, allowing developers to modify available tools without invalidating the prompt cache, and automatic API fallbacks, which can route flagged Opus 5 or Fable 5 requests to another model instead of blocking them.