Anthropic has opened access to Claude Fable 5 for a limited period, giving more paid users a chance to try one of its most capable AI models without paying extra. The move brings a model with Mythos-grade intelligence into everyday use, but only temporarily.
The company is positioning Fable 5 as a general-use model built from the same advanced foundation as Claude Mythos. That matters because Mythos itself was previously kept away from ordinary users due to the risks tied to its security-focused capabilities.
Temporary free access for paid plans
For now, Anthropic is including Claude Fable 5 at no additional cost for Pro, Max, Team, and seat-based Enterprise subscribers. The offer is limited, and the clock is already set.
Access remains free only until 23 June. After that, continued use of Fable 5 will require usage credits, unless Anthropic later decides to restore it as a no-cost benefit inside paid plans.
That makes the rollout less like a permanent upgrade and more like a time-limited test of how widely the model can be used. Users get an early look at Anthropic’s most intelligent consumer-facing option, but the current setup will not stay free forever.
Why Mythos was kept restricted
Claude Mythos was known as a large language model built for highly sensitive security work rather than broad public use. Anthropic viewed wide access as too risky, even though the model was described as stronger than other models in functional terms.
Fable 5 is meant to open much of that power to a wider audience while leaving out the most sensitive parts. Anthropic’s goal is to give users the benefit of its smartest model without exposing capabilities that could be misused.
Built-in safety measures remain in place
Anthropic says Fable 5 still includes safeguards for risky topics. When the system detects a query that falls into those categories, the request is routed to another model instead.
The fallback model is Claude Opus 4.8, which Anthropic describes as its next most capable model. The company says this arrangement helps it launch faster while keeping safety controls active from day one.
Those safeguards are intentionally conservative, which means some harmless requests may also be redirected. Even so, Anthropic says the trigger happens in less than 5% of sessions on average, suggesting the interruption should remain limited for most users.
The company is also acknowledging a broader risk: even without a security-silo focus, a powerful model like Fable 5 could still be tested by bad actors, including for malware-related development. That is part of why the guardrails are built into the rollout.
What the launch says about the AI market
The release reflects how quickly the AI industry is moving toward broader access to more powerful systems. Models that were once considered too sensitive for general users are increasingly being adapted for everyday products.
For Anthropic, Fable 5 is a balancing act between capability and restraint. It expands access to the company’s smartest AI while keeping a clear line around the areas that carry the most risk.
For subscribers on Pro, Max, Team, and Enterprise plans, the immediate benefit is obvious during the promotional window. They can use Mythos-grade intelligence without paying beyond their existing subscription, at least until the free period ends.
After 23 June, the experience changes into a usage-credit model unless Anthropic reverses course. That deadline turns the launch into a short-term opportunity for users who want to evaluate the model while access is still included.
In practical terms, Fable 5 shows how AI companies are starting to release previously restricted capabilities under tighter controls. The model is public enough to be useful, but still guarded enough to reflect the risks Anthropic says cannot be ignored.
