Anthropic's Claude Fable 5 Paves the Way for Public Access to Powerful AI

Anthropic's Fable 5 is a groundbreaking AI model that has been stress-tested for its potential misuse, paving the way for public access to its capabilities.

A Version of Mythos for the Masses

Anthropic has launched Claude Fable 5, the first publicly available version of its Mythos model, which excels at software engineering, knowledge work, and vision. However, the model comes with strict safety limits to prevent high-risk areas like cybersecurity, biology, chemistry, and distillation from being exploited.

The model blocks responses and falls back to Claude Opus 4.8 in these areas, ensuring that even if users attempt to bypass the safety measures, the risks are mitigated. Anthropic has also implemented a 30-day retention policy on all traffic, even for enterprises with zero-retention agreements, to defend against complex and novel attacks.

High-Profile Partners and Early Results

Fable 5 has already garnered attention from various organizations, including Rakuten, which sees the benefits of the model's reflective capabilities in highly autonomous operations. Early testing by Hex, Base44, and Genspark has yielded impressive results, with Fable outperforming other models in complex analytical tasks, UI design, and game coding.

Pricing and Access

While the pricing for Fable 5 is $10 per million input tokens and $50 per million output tokens, double the price of Opus 4.8, the model's capabilities are expected to be in high demand. Access to Fable 5 will be included in Anthropic's Pro, Max, Team, and seat-based Enterprise plans until June 22, after which it will require usage credits.

• Pricing for Fable 5: $10 per million input tokens and $50 per million output tokens
• Access to Fable 5: Included in Pro, Max, Team, and seat-based Enterprise plans until June 22
• Availability: Publicly available, with strict safety limits in place

The Future of AI

Anthropic's move to make Fable 5 available to the public marks a significant step towards making AI more accessible, while also emphasizing the importance of safety measures to prevent misuse. As AI continues to advance, it is crucial that developers and organizations prioritize responsible innovation and adhere to strict safety protocols to mitigate the risks associated with these powerful tools.