Anthropic releases Claude Opus 5, a new heavyweight AI model designed for complex reasoning and efficient coding tasks.

Anthropic has launched Claude Opus 5, a new heavyweight model designed to handle complex, long-running coding tasks with high levels of reasoning and autonomy. The administration announced that the model establishes a new state of the art on coding and knowledge work evaluations, such as Frontier-Bench and GDPval-AA. Notably, Claude Opus 5 outperformed the previous record on the ARC-AGI-3 benchmark, scoring 30.2 percent by solving novel environments through stronger logical reasoning. The model is positioned as a cost-effective alternative to the flagship Fable 5, offering similar performance at roughly half the price. It is designed to be more proactive, capable of verifying its own work and iterating until a task is successful. While it remains behind Mythos 5 in long-running autonomous biology research, it is now the most capable generally available model for scientific research. Additionally, the model features enhanced safeguards that are 85% less restrictive than Fable 5 for cybersecurity tasks, allowing for more efficient vulnerability identification.

Sources