Claude Fable 5 vs Mythos 5: What’s the difference and who gets access?

Anthropic has unveiled a new generation of AI systems called Mythos-class models, introducing what it describes as its most capable family of models so far. The launch marks the first time a Mythos-class model is being made available to the public, but it also comes with an unusual twist: Anthropic is releasing two nearly identical versions of the technology with very different access rules.

The new lineup includes Claude Fable 5, which is available to paying customers, and Claude Mythos 5, a more tightly controlled version reserved for select organisations and security partners.

So what exactly is a Mythos-class model, and why has Anthropic decided to split it into two versions?

What is a Mythos-class model?

Mythos-class is Anthropic's newest category of AI models and succeeds the company's earlier Opus and Sonnet families.

According to Anthropic, these models are designed to handle complex, long-running tasks that require sustained reasoning. The company says Mythos-class systems are particularly strong in software engineering, scientific research, advanced problem-solving and knowledge-intensive workflows where AI must maintain context and reasoning over extended periods.

The Mythos family first appeared earlier this year through a limited-access preview programme, but Anthropic stopped short of releasing it widely due to concerns about some of its capabilities.

Must read: Indian government, tech firm join forces to test Anthropic’s Claude Mythos: Report

What is Claude Fable 5?

Claude Fable 5 is the first Mythos-class model available to the broader public.

Anthropic describes it as its most capable generally accessible AI model to date. The model is designed for developers, researchers and businesses that need AI assistance for coding, analysis, research and complex agentic workflows.

What makes Fable 5 different from previous Claude models is the addition of new safety systems. Anthropic says the model can detect certain categories of high-risk requests and apply additional safeguards before responding.

The company says Fable 5 underwent extensive internal testing, external red-teaming exercises and jailbreak evaluations before release.

What is Claude Mythos 5?

Claude Mythos 5 shares the same underlying architecture as Fable 5 and offers the same core capabilities.

The key difference is that Mythos 5 does not include some of the additional restrictions and safety controls built into Fable 5.

Anthropic has previously indicated that earlier Mythos models demonstrated unusually strong cybersecurity capabilities, including the ability to identify and exploit software vulnerabilities. Those findings led the company to limit access while it evaluated potential risks.

Must read: Anthropic’s Mythos AI raises banking cyber risk concerns, FSB briefing planned

As a result, Mythos 5 remains under tight controls and is not being broadly released.

Why did Anthropic create two versions?

The company appears to be trying to solve one of the biggest challenges facing AI developers today: how to release increasingly powerful models without making potentially dangerous capabilities widely accessible.

Instead of permanently restricting the technology, Anthropic has adopted a tiered approach.

Fable 5 serves as the public-facing model with additional safeguards, while Mythos 5 is reserved for vetted organisations that may require access to the model's full capabilities for research, cybersecurity and infrastructure-related work.

The strategy allows Anthropic to make its latest technology available while maintaining tighter oversight of higher-risk use cases.

What safety measures are built into Fable 5?

One of the most notable additions is a routing system designed for sensitive requests. If users attempt tasks involving advanced cybersecurity, biology or chemistry topics, Fable 5 can either refuse the request or automatically transfer the interaction to the less capable Claude Opus 4.8 model.

Anthropic says this approach helps reduce access to potentially dangerous capabilities without limiting everyday use.

The company also says it carried out adversarial testing, bug bounty programmes and external security reviews before launch. According to Anthropic, testers were unable to identify a universal jailbreak that could reliably bypass the model's safeguards.

Who can use Fable 5?

Fable 5 is available immediately to paying Claude customers, including Pro, Max, Team and Enterprise subscribers. Anthropic is also integrating the model into developer ecosystems and third-party platforms, including GitHub Copilot and cloud services that provide access to Claude models.

Who gets access to Mythos 5?

Mythos 5 remains restricted to a small group of approved organisations. Anthropic says access will primarily be managed through trusted-access programmes and Project Glasswing, an initiative focused on securing critical software infrastructure.

Previous versions of Mythos were made available only to vetted governments, security researchers, infrastructure operators and select organisations. While Anthropic has said access will gradually expand, it has not announced a timeline for a broader release.