Anthropic launches Fable 5 as it prepares for public markets

On Tuesday, Anthropic launched Claude Fable 5, the first publicly available version of its Mythos model. The model excels at software engineering, knowledge work, and vision, but it operates with hard safety limits. In high-risk areas—specifically cybersecurity, biology, chemistry, and distillation—the model blocks responses and falls back to Claude Opus 4.8.

Pricing for the model is:

• $10 per million input tokens
• $50 per million output tokens

To address safety concerns, Anthropic will require a 30-day retention on all traffic to defend against complex and novel attacks, including new jailbreaks, even for enterprises with prior zero-retention agreements. The company noted that there could still be novel attacks, despite testing. According to Anthropic, “Internally, we ran an external bug bounty that produced no universal jailbreaks in over 1,000 hours of testing. We then worked with external red-teaming orgs which also failed to find universal jailbreaks.” The company stress-tested its classifiers with jailbreak attempts before releasing Fable 5.

Early performance data indicates that at least 95% of Fable sessions run entirely on the model’s own responses. Third-party testing has yielded feedback. Analytics company Hex reported that Fable was the first model to score 90% on its core analytics benchmark, stating that on the hardest questions, the model shows strong judgment and attention to nuance. Base44, a vibe-coding platform (a style of coding platform), noted that Fable is better at completing full apps in a single prompt—a process known as one-shotting—and has excellent tool-calling. AI-powered workspace platform Genspark reported that Fable outperformed other models in its evaluations, performing significantly better on tasks like UI design and game coding. Additionally, shopping rewards platform Rakuten stated that at the highest effort, Fable reflects on and validates its own work.

The deployment expands access across 15 countries. Mythos was initially launched as a preview in April, limited to a handful of partners due to cybersecurity concerns, before Anthropic expanded access to hundreds of organizations last week. The launch arrives as Anthropic prepares to enter the public markets, alongside peers like OpenAI and Elon Musk’s SpaceX. Anthropic has also warned that AI systems are advancing so rapidly that they may soon achieve recursive self-improvement (RSI), autonomously improving themselves without human intervention.

Why it matters

Anthropic is bringing its most powerful AI model to the general public for the first time, while simultaneously preparing to enter the public markets.