Anthropic has announced the launch of Claude Fable 5 and Claude Mythos 5, two new artificial intelligence models that represent an important step forward for the Claude family.

The main highlight is Claude Fable 5, described by the company as a Mythos-class model designed for general use. In practice, this means it brings much more advanced capabilities, while still including safety layers to help prevent dangerous uses.

According to Anthropic, Fable 5 outperforms all previous Claude models that have been made publicly available. It stands out especially in long, complex tasks that require continuous reasoning.

What makes Claude Fable 5 different?

Claude Fable 5 was built to handle work that requires autonomy, deep analysis, and multi-step execution.

Its main strengths include:

Software engineering
The model can work with large codebases, understand complex structures, and make broad changes more efficiently. In early tests, companies reported that tasks that could have taken weeks or months were completed in just a few days.

Knowledge work
Fable 5 also shows improvements in document analysis, table interpretation, chart reading, financial report understanding, and complex problem-solving. This makes the model useful for areas such as business, research, finance, and operations.

Vision
Another strong point is visual understanding. The model can interpret images, extract data from scientific figures, and even rebuild applications from screenshots. This expands its use in design, engineering, visual analysis, and interface development.

Long context and memory
Fable 5 can stay focused on very long tasks, working with millions of tokens and using its own notes to improve results throughout the process. This is important for large projects, detailed research, and workflows that cannot be completed in a single response.

What about Claude Mythos 5?

Claude Mythos 5 uses the same base model as Fable 5, but with some protections removed in specific areas. Because of this, it will not be available to everyone.

At first, Mythos 5 will be used by selected groups, such as cyber defenders, critical infrastructure providers, and later biomedical researchers through trusted access programs.

The reason for this restriction is simple: models with this level of capability can be extremely useful for protecting systems, finding vulnerabilities, and accelerating scientific research. However, if used incorrectly, they could also make cyberattacks or dangerous biology and chemistry applications easier.

Safety as a central part of the launch

Anthropic made it clear that the release of Fable 5 comes with new safety protections.

When the system detects requests related to sensitive topics, such as offensive cybersecurity, biology, chemistry, or attempts to extract the model’s advanced capabilities, the response may be routed to Claude Opus 4.8.

This means that in some cases, the user may not receive a direct response from Fable 5. The company says this choice was made to release the model more quickly and safely, even though some legitimate requests may be blocked by mistake.

According to Anthropic, more than 95% of sessions should not trigger this fallback. Still, the company recognizes that the filters may be too conservative at first and plans to adjust them over time.

Applications in science and research

One of the most interesting parts of the announcement is the use of Mythos 5 in scientific research.

Anthropic says the model helped internal experts in protein design speed up parts of the drug discovery process. In some tests, Mythos 5 was able to carry out steps normally performed by scientists, such as choosing targets, selecting tools, running analyses, and correcting failures during the process.

The model also showed the ability to generate scientific hypotheses in molecular biology. For Anthropic, this kind of progress shows how more powerful models can help researchers find new paths in complex fields.

Pricing and availability

Claude Fable 5 is available for general use, while Claude Mythos 5 remains restricted to selected partners.

The announced price for both models is $10 per million input tokens and $50 per million output tokens.

For users on Pro, Max, Team, and Enterprise per-seat plans, Fable 5 will be temporarily included at no extra cost until June 22, 2026. After that, usage will require credits until Anthropic can expand capacity and possibly bring the model back into subscription plans.

What does this launch mean?

The launch of Claude Fable 5 shows that AI models are entering a more advanced phase: less focused only on answering simple questions and more focused on executing long, technical, and autonomous tasks.

For developers, this may mean greater productivity in large projects. For companies, it may bring deeper analysis and smarter workflows. For researchers, it may open new possibilities in science, biology, and engineering.

At the same time, Anthropic reinforces that the more powerful AI becomes, the greater the need for safety, control, and responsible access.

Claude Fable 5 appears to be an important step in that direction: a more capable, more autonomous, and more useful model, but released with clear limits to reduce risks.

Mythos 5 shows the other side of this new generation: extremely strong capabilities, but available only in trusted contexts where the positive impact can be greater and the risks can be better controlled.