Archive — US — techcrunch
Anthropic’s new Fable release is cheaper, less restrictive
Anthropic has released new AI models, Fable and Mythos 5.1, which are cheaper and less restrictive compared to earlier versions. Fable will be accessible to a wider audience, while Mythos remains limited to partners in cybersecurity and life sciences. Notable features include zero data retention and a high-privacy monitoring service, alongside performance upgrades and improved benchmarking results.
Key points
- Released Fable and Mythos 5.1, advancing AI capabilities with performance upgrades.
- Fable 5.1 is unrestricted and available on cloud platforms, while Mythos 5.1 is restricted to registered partners.
- Introduced zero data retention, allowing users to run models on their infrastructure securely.
- New monitoring service, Enterprise Frontier Safeguards, rolling out in the fall allows clients to control misuse monitoring.
- Models achieved record scores on benchmarks like Terminal-Bench 4.0 and Humanity’s Last Exam.
- The system card rates Mythos as 'low-risk' concerning self-improvement capabilities.
Why it matters
- The trend toward zero data retention and privacy-focused AI services aligns with increasing concerns surrounding data security in AI deployments.
- 2025 IT trends indicate a push for more ethical AI practices and enhanced user control over AI, reflecting concerns around AI misuse and data privacy.
- Anthropic's commitment to not training on enterprise data without permission mirrors growing regulations and ethical standards around AI data use.
The release of Fable and Mythos 5.1 reflects Anthropic's commitment to improving AI accessibility and privacy, aligning with broader trends in the tech industry toward ethical AI development and security.