Anthropic released Claude Fable 5, its first Mythos-class model, which achieves record benchmark performance but doubles the price of previous versions. The model also implements stricter safety filters and new data retention policies.
Anthropic has launched Claude Fable 5, marking the debut of its new Mythos model tier. The system dominates industry benchmarks, achieving 95 percent on SWE-bench Verified and leading nearly every other major performance metric.
Pricing represents a significant shift. Claude Fable 5 costs $10 or $50 per million tokens, doubling the expense of Claude Opus 4.8. This positions it as Anthropic's premium offering, targeting enterprises and organizations requiring maximum performance.
Safety measures have intensified. The model's built-in filters block approximately nine percent of user requests, a stricter threshold than previous versions. Anthropic has not specified which request types trigger the blocks, citing safety concerns.
A new 30-day data retention policy now applies universally. Previously, customers could purchase zero-data-retention contracts, allowing immediate deletion of conversation data. This change affects all users, including enterprise customers who previously maintained these agreements. Anthropic has not announced exceptions to this policy.
The Mythos classification signals Anthropic's tiered approach to AI development. This class represents models built with enhanced capabilities and safety measures, distinguishing them from previous release cycles. Industry observers view this as a positioning strategy against competitors offering similar performance tiers.
The benchmark performance metrics demonstrate substantial improvements in coding tasks, where the 95 percent SWE-bench score reflects advanced software engineering capabilities. This positions Claude Fable 5 competitively within the market for AI-assisted development tools.
Customers upgrading from previous Claude versions face cost and operational considerations. The doubled pricing requires budget adjustments, while the universal data retention policy may conflict with existing data governance frameworks at some organizations.
Anthropić has not announced a timeline for additional Mythos-class models or whether pricing will adjust based on demand.
Jacob Tsimerman, a newly awarded Fields Medalist, is leaving the University of Toronto to join OpenAI's safety research efforts. The mathematician recently published research analyzing how AI systems could contribute to human extinction scenarios.
AI agents dramatically outpace simple chatbots in energy consumption, according to climate scientist Zeke Hausfather's eight-week analysis of Claude Code usage. His findings reveal a stark gap between reported energy figures and actual operational costs.
xAI has released Imagine Image 2.0, a new image generator integrated into Grok that scores second in Arena benchmarks, trailing only OpenAI's GPT-Image-2. The model includes new editing tools and workflow templates designed for practical creative use.
The U.S. Department of Energy has announced the Genesis Open Models Initiative, a program aimed at developing and democratizing artificial intelligence models for scientific research and industrial applications.