Anthropic Releases Claude Fable 5, Public Version of Mythos Model with Safety Guardrails
Anthropic launched Claude Fable 5 on Tuesday, making its powerful Mythos-class AI model publicly available for the first time through its API and subscription plans, though with safety restrictions that defer to a less capable model for high-risk queries. The model was initially restricted to select partners in April due to cybersecurity concerns, but is now being rolled out more broadly with mandatory data retention policies and higher pricing. The release represents a significant step in making frontier AI capabilities accessible while attempting to balance capability with safety measures.
Anthropic announced the public availability of Claude Fable 5, a version of its Mythos-class model that represents the company's most powerful publicly released AI system. The model excels at software engineering, knowledge work, vision, and scientific research, but includes safety guardrails that automatically defer to the less capable Claude Opus 4.8 for high-risk domains including cybersecurity, biology, chemistry, and distillation. Fable 5 is available immediately through June 22 at no extra cost on Pro, Max, Team, and Enterprise subscription plans, after which it will require usage credits. Anthropic simultaneously deployed Mythos 5 to pre-approved enterprise partners. Both models are priced at $10 per million input tokens and $50 per million output tokens—double the cost of Opus 4.8 but significantly cheaper than the earlier Mythos Preview. The company implemented a new mandatory 30-day data retention policy for all Mythos-class model traffic, even for organizations with zero-retention agreements, citing the need to defend against novel attacks and jailbreaks, though it claims the data will not be used for model training.
What's missing
Neither source adequately addresses the competitive landscape or how Fable 5 compares to other frontier models like OpenAI's latest offerings or Google's Gemini 3.5 Flash (which The Register notes was omitted from Anthropic's benchmark comparisons). Additionally, there is limited discussion of the practical implications of the 30-day data retention requirement for enterprises with strict privacy needs or regulated industries.
How coverage differed
Both sources report the same core facts neutrally, though The Register adopts a slightly more skeptical tone with phrases like 'spins a Fable' and 'wreak havoc,' while TechCrunch emphasizes Anthropic's safety testing and industry-leading performance benchmarks more prominently. The Register also highlights the data retention policy change more critically as affecting organizations with 'zero data retention' policies.
What different sources said
- The RegisterCenter
Anthropic spins a Fable of a tamer, safer Mythos
- TechCrunchCenter
Anthropic’s Claude Fable 5 is a version of Mythos the public can access today
Related
Apple's Upgraded Siri AI Shows Practical Improvements in Real-World Testing
Apple has released an upgraded version of Siri with improved AI capabilities that can now handle practical tasks like extracting event information from emails and creating calendar entries. The new Siri can also assist with tasks like creating shopping lists, setting reminders, and providing gardening advice by referencing user information. The improvements address long-standing limitations of the original AI-enhanced Siri and demonstrate Apple's effort to make voice assistants more useful for everyday household management.
Hardware Integration Becomes Focus of Modern Hackathons as Software Development Becomes Commoditized
A hackathon participant describes how modern hackathons are shifting focus from pure software development to hardware integration projects, exemplified by their team's AI-powered rotary phone that plays music via Spotify. This shift reflects how AI tools and no-code platforms have made traditional software development faster and less challenging, freeing participants to focus on physical hardware interfaces. The trend suggests hackathons may increasingly emphasize creative hardware projects over conventional software applications as the bar for impressive software-only projects continues to lower.
Taiwan Considers Stricter AI Chip Export Controls to China to Align with U.S. Restrictions
Taiwan is considering implementing stricter export controls on AI chip sales to China to better align with U.S. semiconductor restrictions and combat smuggling. Currently, Taiwan lacks specific laws treating unauthorized AI chip exports to China as crimes, relying instead on enforcement through other existing regulations. This move matters because it would close legal gaps that allow advanced semiconductor diversion and strengthen the U.S.-Taiwan-led effort to prevent China from accessing cutting-edge AI technology for military purposes.