Anthropic is giving Claude another major update today with the launch of Fable 5.1, the newest version of its flagship Mythos-class artificial intelligence model. The release follows Fable 5, which arrived in June, and it promises improvements across coding, knowledge work, and long-running problem-solving tasks. Fable 5.1 is available now, alongside Claude Mythos 5.1 for users who participate in Anthropic's trusted access programs.

Anthropic says Fable 5.1 and Mythos 5.1 are the same model underneath, with different safeguard levels. Fable is generally available, while Mythos is limited to selected users. The difference is central to how Anthropic plans to support work in cybersecurity and the life sciences. It lets the company pair its most advanced model with stricter controls when necessary.

Claude Fable 5.1 and Mythos 5.1 arrive today

In Anthropic's model hierarchy, Fable sits at the very top. Haiku, Sonnet, and Opus have long served as the small, medium, and large models in the Claude lineup, but Fable is positioned as the most capable option. Fable 5.1 builds on that by adding improvements for the kinds of work enterprises find hardest to automate: complex coding, deep research, and tasks that require a model to keep context alive over many steps.

The launch also makes Fable 5.1 available through public APIs and Anthropic's products. Although no further details were given in the initial release, the company says the new model sets a new standard on coding, knowledge work, and long-running problem solving. Anthropic also says Fable 5.1 achieves similar or better results than Fable 5 at a much lower cost when set to low or medium effort.

Fable and the rest of the Claude lineup

For anyone not following Anthropic's model naming closely, it is useful to separate the layers. Claude's best-known tiers are Haiku, Sonnet, and Opus. Haiku is the small model, built for fast and inexpensive responses. Sonnet is the middle option, intended for a broad set of everyday tasks. Opus is the large model, designed for advanced reasoning and more open-ended work. Fable is now the upper tier, above Opus, and it represents Anthropic's most ambitious approach to agentic workflows.

With Fable 5.1, Anthropic is not simply changing a version number. The company says the new model has a higher level of performance compared to Fable 5 while also being more efficient in many situations. The fact that Fable 5.1 can return results that are similar or better than Fable 5 at a lower cost in certain effort settings means developers do not always need to select the highest possible compute mode. Instead, they can choose where high effort is worth the extra price and where lower effort is enough.

What Fable 5.1 brings to coding and knowledge work

One of the defining features of Fable 5.1, according to Anthropic, is its behavior when solving software problems. The company says Fable 5.1 avoids shortcuts that result in poorer-quality work, and it is smart enough to fix root causes of issues rather than simply layering workarounds on top. In testing by the investment firm Millennium, for example, Fable 5.1 found the cause of a rare crash in internal systems that engineers and other models had been unable to explain after several years. Such a result is significant because many coding assistants can suggest plausible patches but struggle when failures happen only under unusual conditions.

For knowledge workers, Fable 5.1 is designed to handle tasks that involve research, synthesis, planning, and report generation. Long-running problem solving is another major focus. Many enterprise workflows require an AI assistant to work through many steps, read new information along the way, and return to earlier decisions when the situation changes. According to Anthropic, Fable 5.1 offers improvements in this area as well.

Mythos 5.1 arrives with tailored safeguards

The second release today is Claude Mythos 5.1. Anthropic says Mythos is available only through trusted access programs, and its safeguards are specifically designed to support work in cybersecurity and the life sciences. This does not mean Fable 5.1 is a different model from Mythos 5.1. Both versions share the same underlying capabilities. The distinction is in how the systems are gated and protected for different use cases.

In cybersecurity, organizations often handle large code bases, threat data, and incident logs that require careful access controls. In the life sciences, research teams may work with sensitive datasets, biomedical literature, and regulated information. Providing a tailored version of the model makes it easier for Anthropic to offer advanced AI in these contexts without loosening the safety measures needed outside a general consumer or business product. Users with access to Mythos will see a system intended for high-trust environments rather than a consumer chatbot experience.

Pricing and availability

Pricing is one area where Fable 5.1 differs from its predecessor in concrete terms. Anthropic says Fable 5.1 will cost an estimated 25 percent less than Fable 5 for typical workloads, wherever usage is billed by token. That saving comes from reduced pricing on cache reads, which occur when the model reads inputs that have already been processed and stored. In applications where the same context is used repeatedly, cache-read price reductions can be substantial.

Anthropic says the company has reduced the price of Fable 5.1's cache reads in token-based billing systems such as the API. This leads to a significant reduction in the overall cost of running the model. For highly agentic work, the savings will often be much larger, up to approximately 45 percent. Pricing is otherwise unchanged from Fable 5. That means most of the cost advantage comes from cache economics rather than from simply lowering every aspect of the model's price.

For developers, the change is especially useful in coding assistants and virtual agents that keep conversation contexts alive over many turns. A model that can remember earlier code, user messages, and tool outputs is powerful, but it also adds more tokens to every request. Cutting the cost of reading cached tokens makes it less expensive to maintain long sessions and to run models across large repositories or documents.

Why the update arrives so quickly

The timing of Fable 5.1 is notable. Fable 5 was first released in June, meaning today's upgrade comes only three months later. In previous model cycles, major model versions have sometimes taken much longer to reach customers. A faster update window suggests Anthropic is iterating aggressively and sees

Source: 9to5Mac News