Anthropic launches Claude Opus 5, a new artificial intelligence model offering high-level performance at half the cost of its most powerful counterpart.
Anthropic released Claude Opus 5 on Friday, a new artificial intelligence model designed to provide near-frontier intelligence at a significantly lower price point. The San Francisco-based startup stated that the model offers performance close to Claude Fable 5 while costing half as much per task. This release addresses growing concerns among business customers regarding the high costs of running advanced AI systems. The model excels in coding and knowledge work, topping several industry benchmarks such as Frontier-Bench v0.1. While it remains behind the secretive Mythos 5 model in complex cybersecurity tasks, Anthropic designed Opus 5 to be more efficient and less restrictive. It features improved safeguards that are 85% less likely to intervene than those on Fable 5. Additionally, the model is free from the 30-day data retention policy that affected previous versions, making it more appealing to privacy-conscious users. Anthropic also introduced an Automatic Fallbacks feature to ensure users receive functional responses when safety classifiers are triggered.