Anthropic launches Opus 5 with stronger safeguards, near-Fable 5 power
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic released Claude Opus 5, a new AI model that comes close to Fable 5’s capabilities in many domains and shows significant improvement in complex coding tasks
- Anthropic added stronger cybersecurity safeguards to Opus 5 than were present in Opus 4.8, following government concerns over Fable 5 and a subsequent regulatory review period
- Anthropic confirmed it is working with government partners to conduct independent testing of Opus 5, though it did not disclose whether the model was pre-approved by the Trump administration
- Anthropic marketed Opus 5 as the ideal everyday model for enterprise customers, especially in knowledge work and biology, while positioning Fable 5 for more ambitious, long-horizon AI agent projects
- Anthropic maintained the same pricing for Opus 5 as its predecessor—$5 per million input tokens and $25 per million output tokens—and emphasized it is half the price of Fable 5 and slightly cheaper than OpenAI’s GPT-5.6
- Anthropic introduced a 'Fast mode' for Opus 5 in research preview, offering higher speeds at double the cost, and added an option to automatically fall back to a lower-tier model when safeguards block a request
Why it matters: Enterprises gain access to a more secure, cost-efficient model that narrows the performance gap with top-tier AI systems, while Anthropic signals compliance with emerging regulatory expectations—potentially shaping industry norms as OpenAI and others face similar scrutiny. The pricing strategy counters concerns about AI cost pass-through, directly addressing customer friction.
