Anthropic unveiled Claude Opus 5 on Thursday, marking the company’s latest product release amid ongoing regulatory scrutiny. The new model demonstrates performance levels comparable to Claude Fable 5 across numerous areas and shows significant improvements in handling complex coding assignments. This launch arrives weeks after government negotiations regarding Fable 5 and shortly following a notable security incident at rival OpenAI that captured industry attention.
The company emphasized that Opus 5 incorporates enhanced cybersecurity protections beyond those found in its predecessor, positioning it as the most aligned Opus variant to date and most resistant to misuse attempts. Anthropic indicated the model serves enterprise customers best for knowledge work and biological applications, while reserving Fable 5 for more demanding tasks and long-term AI agent initiatives. A spokesperson confirmed that Anthropic continues collaborating with government partners who conduct independent testing of the company’s models, including Opus 5.
Pricing for Opus 5 matches its predecessor at five dollars per million input tokens and twenty-five dollars per million output tokens, undercutting Fable 5 by half and remaining competitive with OpenAI’s comparable offering. The company introduced a “Fast mode” option launching in research preview that doubles processing speed for additional cost, alongside an automatic fallback feature allowing users to request lower-tier models when safeguards restrict requests.
Originally reported by Theverge →
About the Author
