Anthropic releases Opus 5 promising Fable 5-like capabilities

Key Points:

• Anthropic has launched Claude Opus 5, a new AI model that rivals the capabilities of Claude Fable 5 in many areas and excels in complex coding tasks, while incorporating stronger cybersecurity measures than its predecessor, Opus 4.8.
• The release follows recent regulatory scrutiny and negotiations with the US government, with Anthropic confirming ongoing independent government testing of Opus 5 prior to its public rollout.
• Opus 5 is targeted primarily at enterprise customers for knowledge work and biology, positioned as an everyday model, whereas Fable 5 remains suited for more ambitious AI projects.
• Pricing for Opus 5 remains consistent with the previous model and is half the cost of Fable 5, addressing concerns about hidden usage caps and rising AI costs; a "Fast mode" option is available for faster responses at double the price.
• The model also offers an automatic fallback feature to a lower-tier model when usage safeguards restrict certain requests, enhancing alignment and reducing susceptibility to misuse.