Anthropic has pulled its two most capable models from service. On the evening of June 12, 2026, the company suspended access to Claude Fable 5 and Claude Mythos 5 after receiving a directive from the United States government that cited national security concerns.
A Sudden Suspension
According to Anthropic, the order arrived at 5:21pm ET on June 12 and took effect immediately. The suspension is unusually broad. It applies to everyone, including commercial customers, individual users, and even Anthropic employees working outside the United States. For a model that was reaching hundreds of millions of people, an instant global cutoff is close to unprecedented.
What Is Affected, and What Is Not
The directive names only two models. Fable 5 was Anthropic's newest flagship, sitting at the very top of the lineup, while Mythos 5 is the more capable of the pair and had been available to a narrower set of users. Both are now offline.
Every other Claude model remains fully available. That includes Claude Opus 4.8, Opus 4.7, Opus 4.6, Sonnet 4.6, and Haiku 4.5. If your application runs on any model other than the two named above, nothing changes for you today.
The Reason Given
The government order referenced a potential jailbreak method uncovered during testing. Anthropic's own review describes the issue in narrow terms. It involves asking the model to read a specific codebase and then fix any software flaws it finds. The company points out that this kind of capability is widely available from other frontier models, including OpenAI's GPT-5.5, and is not something unique to Claude.
Anthropic Pushes Back
The company made clear that it does not agree with the decision. In its statement, Anthropic wrote that it disagrees "that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people."
Anthropic pointed to its defense in depth strategy, a layered set of safeguards meant to catch misuse before it causes harm, and raised a broader concern. Applying this standard across the whole industry, the company suggested, would essentially halt all new model deployments. The argument is straightforward: if every narrow jailbreak triggered a recall, no frontier model from any lab would stay on the market for long.
What This Means If You Built on Fable 5
Fable 5 launched only weeks ago as the new top of Anthropic's range, priced at $10 per million input tokens and $50 per million output tokens. Teams that adopted it for long-horizon coding, complex refactors, or deep research now need a fallback, at least until the situation is resolved.
The encouraging part is that the most natural replacement is also cheaper. Claude Opus 4.8 shares the same request surface as Fable 5, uses the same adaptive thinking mode, and carries the same 1M-token context window, all at half the price. For most workloads it is a near drop-in switch.
| Model | Input $/1M | Output $/1M | Context | Status |
|---|---|---|---|---|
| Claude Fable 5 | $10.00 | $50.00 | 1M | Suspended |
| Claude Mythos 5 | n/a | n/a | n/a | Suspended |
| Claude Opus 4.8 | $5.00 | $25.00 | 1M | Available (drop-in) |
| Claude Sonnet 4.6 | $3.00 | $15.00 | 1M | Available (budget) |
How Long Will This Last?
Anthropic has not offered a timeline, and the outcome rests on a government process rather than a product roadmap. For planning purposes, treat the suspension as open-ended. If Fable 5 is central to your stack, move critical traffic to Opus 4.8 now and keep the more expensive model as an option for later rather than a dependency.
We will keep this page updated as the situation develops. Our live pricing table reflects current availability, you can line up the alternatives side by side in the comparison tool, and the cost calculator will show exactly what switching to Opus 4.8 does to your monthly bill. For background on the model itself, see our earlier Claude Fable 5 pricing breakdown.