Anthropic Claude Fable 5 Launches as State-of-the-Art AI Model Beating GPT-5.5

Anthropic released Claude Fable 5 today as a new general-purpose AI model that competes directly with OpenAI's GPT-5.5. The company also launched Claude Mythos 5 alongside the release, which targets specialized cybersecurity and biology use cases through Project Glasswing. This dual launch expands Anthropic's offerings for both everyday users and enterprise developers.

Anthropic releases dual models with competitive pricing and strict data retention rules

Claude Fable 5 serves as the standard model for general tasks while Claude Mythos 5 provides enhanced capabilities for technical domains. Both models operate on the same underlying architecture but differ in their safety configurations and target applications. The release brings these tools to existing Anthropic platforms including API endpoints and subscription tiers.

Anthropic sets pricing at $10 per million input tokens and $50 per million output tokens for both models. The company enforces a mandatory 30-day data retention policy for all business traffic on first- and third-party surfaces. This retention period applies to Claude Fable 5 and Claude Mythos 5 without exception.

Anthropic claims the new models outperform OpenAI's GPT-5.5 across software engineering, knowledge work, vision, scientific research, and autonomous tasks. The company describes Claude Fable 5 as state-of-the-art for vision tasks while noting Claude Mythos 5 holds the strongest cybersecurity capabilities of any model globally. Anthropic reports that safety safeguards trigger in less than five percent of sessions on average.

Claude Fable 5 is available today through the Claude API and consumption-based Enterprise plans. The model remains included at no extra cost for Pro, Max, Team, and seat-based Enterprise customers from today through June 22. The listed rates apply globally with no region-specific variations.

Anthropic positions this release as a direct challenge to OpenAI's recent offerings in the competitive AI market. Unlike earlier releases that prioritized a balance between safety and utility, this update highlights superior performance metrics in diverse fields such as software engineering and scientific research. The company maintains that it will not use retained business data to train future Claude models.