Anthropic has unveiled Claude Opus 5, an advancement over its predecessor, Opus 4.8. Designed for everyday applications, Opus 5 delivers performance nearly on par with the flagship Fable 5 model but at a significantly reduced cost.
Opus 5 demonstrates superior capabilities across various domains, including agentic coding, innovative problem-solving, knowledge work, computer utilization, and multidisciplinary reasoning. Notably, it offers substantial improvements in scientific research tasks and generates enhanced visual outputs compared to Opus 4.8.
Despite its advancements, Opus 5 maintains the same pricing structure as Opus 4.8, at $5 per million input tokens and $25 per million output tokens. Additionally, a Fast mode is available, operating at 2.5 times the standard speed for double the base price.
While Opus 5 approaches the proficiency of Mythos 5 in identifying cybersecurity vulnerabilities, it was not specifically trained for cyber tasks and remains considerably less adept at exploitation. Mythos 5, essentially Fable 5 with certain safeguards removed, is restricted to a select group of organizations, primarily cyber defenders and infrastructure providers, due to its advanced capabilities in vulnerability detection and exploitation.
Opus 5 is now accessible across all platforms, serving as the default model for Claude Max and representing the most robust option available for Claude Pro users.
The introduction of Claude Opus 5 signifies a strategic move by Anthropic to democratize access to high-performance AI models. By offering capabilities close to their top-tier Fable 5 at a more affordable price point, Anthropic is positioning itself as a formidable competitor in the AI landscape, potentially influencing pricing and accessibility standards industry-wide.
