Anthropic has extended paid subscriber access to Fable 5 through July 12, while simultaneously launching Claude Sonnet 5 as a more affordable and safer alternative following government intervention that temporarily halted its most powerful models. The two developments reflect the company's effort to navigate heightened regulatory scrutiny while maintaining momentum in the competitive AI market.

What Happened to Fable 5 and Why Was Access Extended?

Fable 5, described as a "defanged" version of Anthropic's more powerful Mythos model, was pulled from public access on government orders just four days after its release. The model was restricted from responding to high-risk queries about topics like cybersecurity and biological weapons, yet still represented a significant capability jump.

The temporary shutdown created complications. Safety testers discovered that Fable 5 was set to downgrade to Opus in response to certain questions without their knowledge, creating trust issues between researchers and Anthropic. Amazon researchers later jailbroke Fable 5 and alerted the White House about the vulnerability, though Anthropic characterized the incident as "narrow" in scope.

On June 30, the Department of Commerce lifted export controls on both Mythos 5 and Fable 5. Anthropic began restoring global access to Fable 5 on July 1. However, Anthropic had originally set Fable 5 to shift to token-based usage on July 7 for paid users. The company has now extended paid plan access to July 12, giving subscribers additional time before pricing changes take effect.

"As before, you can use up to 50% of your weekly usage limit on Claude Fable 5. After that, you can keep using Fable 5 with usage credits, or switch to another model to keep working within your remaining limits," Anthropic explained in a statement.

Anthropic, official statement

How Does Claude Sonnet 5 Differ From Previous Models?

Anthropic released Sonnet 5 as a significant step forward in its Claude lineup, positioning it as a model that can "make plans, use tools like browsers and terminals, and run autonomously at a level that, just a few months ago, required larger and more expensive models." The company stated that Sonnet 5 performs similarly to its more expensive Opus 4.8 model but at substantially lower cost.

Sonnet 5 starts at $2 per million input tokens but will increase to $3 per million in September. It is now the default in Free and Pro plans and available to all other plan types, including Max, Team, and Enterprise. Testing revealed that Sonnet 5 scores notably high on computer use benchmarks and agentic coding tasks, completing complex tasks that earlier Sonnet models could not handle.

The timing of Sonnet 5's release reflects a deliberate strategy around safety messaging. Anthropic emphasized that Sonnet 5 "has a much lower ability to perform dangerous cybersecurity tasks than our current Opus models," a notable distinction following the complex rollout of more powerful Fable 5 and Mythos 5 models. The model comes with safeguards implemented automatically, reflecting both recent public relations challenges and broader improvements in general intelligence.

Anthropic

How to Understand Anthropic's Layered Model Strategy

• Capability Positioning: Sonnet 5 serves as a mid-tier model balancing performance and cost, positioned between less capable models and more powerful Opus variants, which remain restricted or under government oversight.
• Safety Emphasis: Anthropic deliberately highlights that Sonnet 5 has lower ability to perform dangerous cybersecurity tasks than Opus models, a messaging choice reflecting government intervention and public scrutiny of more powerful models.
• Pricing Strategy: Sonnet 5 starts at $2 per million input tokens before rising to $3 per million in September, making it significantly cheaper than Opus models while claiming near-flagship performance for many use cases.
• Access Flexibility: The July 12 extension for Fable 5 paid access allows subscribers to evaluate the model before token-based pricing takes effect, while Sonnet 5 is immediately available across all plan types.

What Does Government Intervention Signal About AI Regulation?

The situation surrounding Fable 5, Mythos 5, and Sonnet 5 represents a rare instance of U.S. authorities intervening in an American AI product in public. This could signal a shift in the relatively hands-off approach to AI labs that has characterized recent policy, particularly under the Trump administration.

Mythos 5 remains available only to specific institutions, indicating that government concerns about the most powerful models persist. The Department of Commerce's decision to lift export controls on both models suggests a more nuanced regulatory approach, where capability levels and intended use cases determine access rather than blanket restrictions.

For enterprises and developers, Sonnet 5's positioning as a capable yet safer alternative may appeal to organizations concerned about regulatory risk. The model's availability across all plan types and its default status in Free and Pro plans suggest Anthropic is betting on broad adoption to offset restrictions on its more powerful models. The extended access period through July 12 for Fable 5 gives paid subscribers time to evaluate both models before making long-term decisions about their AI infrastructure.