For years, the most capable Claude model has carried the name Opus. In 2026, Anthropic introduced a new tier above it, called Mythos. Shortly after the first public release in that tier, access was paused for nearly three weeks because of a US government directive. If you are trying to make sense of the names and the news, this guide lays out what is known.
The Claude model tiers
Anthropic sells Claude in tiers of increasing capability and cost. Haiku is the smallest and fastest. Sonnet is the balanced everyday model. Opus has long been the most capable of the three.
Mythos is a newer tier above Opus. Anthropic's first model in this class, Claude Mythos Preview, was not released to the general public. It has been used by a small number of trusted organisations as part of an Anthropic initiative called Project Glasswing.
Mythos and Fable: two names, one model
The current generation in this tier has two names: Claude Mythos 5.1 and Claude Fable 5.1. This can be confusing, so it helps to be precise.
According to Anthropic, Mythos and Fable share the same underlying model. The difference is that Fable has additional safety measures in three sensitive areas: biology, cybersecurity and research on large language models themselves. In other words, Fable is the version with extra guardrails, built for wider availability.
Claude Fable 5.1 is among the publicly available models, alongside Opus 5.5, Sonnet 5.5 and Haiku 5.5.
Why the extra safeguards
As AI models become more capable, the risks of misuse in certain areas grow too. Biology is a concern because detailed technical help could, in the wrong hands, contribute to dangerous outcomes. Cybersecurity is a concern because a highly capable model could help attackers find and exploit weaknesses. Research on AI systems is sensitive because it could speed up the development of even more powerful systems without adequate oversight.
Anthropic has published a Responsible Scaling Policy that commits it to stronger safeguards as its models become more capable. The split between Mythos and Fable reflects that approach: the most capable version is restricted, and the widely available version adds protections in the highest-risk areas.
What happened in June 2026
Anthropic first released Claude Fable 5 and Claude Mythos 5 on 9 June 2026. On 12 June, Anthropic suspended access to both models to comply with export controls issued by the US Department of Commerce.
The Department of Commerce lifted those controls on 30 June 2026, and Anthropic restored access on 1 July 2026. Anthropic published a statement about the episode, which is the best primary source for details: anthropic.com/news/fable-mythos-access.
During the suspension, users who had started relying on the new models had to switch back to other Claude models.
Why export controls apply to AI at all
Export controls are government rules that restrict how certain technologies can be sold or shared, usually for national security reasons. The United States has used them for many years on items such as advanced computer chips, encryption technology and military equipment.
Recently, governments have started treating the most advanced AI systems and the chips that run them as strategically important technologies. That is why export rules increasingly affect not only hardware but also the AI models themselves.
The details of why the controls were applied in June and then lifted are a matter of US government policy. Different observers have taken different views on whether such controls are the right tool for managing AI risks, and the debate is ongoing.
What it means for users and businesses
For individuals using Claude through the app, the main lesson is that the most advanced AI models now sit at the intersection of technology and government policy. Availability can change for reasons that have nothing to do with the technology itself.
For businesses building products on top of AI models, the episode is a reminder to plan for change:
- Avoid depending entirely on a single model. Design systems so you can switch to another model if one becomes unavailable.
- Test fallback models in advance so a switch does not break your product.
- Watch official announcements from AI providers, which tend to explain availability changes faster and more accurately than social media.
- For companies outside the United States, including in India, keep an eye on how international rules evolve, since they can affect access to the most advanced models.
Should you use Fable 5.1?
For most everyday tasks, such as writing, summarising, analysis and routine coding, Sonnet 5.5 or Opus 5.5 are more than capable. Fable 5.1 is aimed at the hardest problems, where its extra capability justifies its higher cost. As always, the best way to decide is to test the models on your own work and compare quality, speed and cost.
The bottom line
The Mythos tier is Anthropic's top level of capability, with Mythos and Fable sharing the same model and Fable adding extra safety measures for wider release. The June 2026 pause showed how closely advanced AI is now tied to government policy. For users, the practical takeaway is simple: understand the tiers, pick the model that fits the task, and build in flexibility in case availability changes.
