Claude Fable 5 is available again. Anthropic said the U.S. export controls were lifted on June 30, and that Fable 5 would return on July 1 across Claude.ai, Claude Code, Claude Cowork, and the developer-facing Claude Platform. A June 12 government order had shut it down entirely — 19 days earlier.
Pro, Max, Team, and select Enterprise users get Fable 5 included in their plans only through July 7; after that, they have to turn on a separate metered charge called usage credits. Three clouds — AWS, Google Cloud, and Microsoft Foundry — and the restricted Mythos 5 model are on different footing again.
Where it's available
Anthropic said Fable 5 is available "to users globally," but "globally" here means the countries the company serves — a defined supported-countries list. Places where Claude isn't offered at all, whether under sanctions or other restrictions, aren't on it. South Korea, by contrast, is on that list for both Claude.ai and the API, so users there can use Fable 5.
After July 7, it's usage credits
Through July 7, the Pro, Max, Team, and select Enterprise plans include Fable 5 on a temporary basis, for up to 50% of weekly usage limits. Anthropic hasn't published this as an absolute message or token cap, so the real limit varies by plan and usage.
After July 7, Fable 5 moves to usage credits — a metered charge that sits alongside the subscription. Once the included usage is spent, continuing means turning on credits and paying for the extra at standard API rates. On individual plans (Pro, Max 5x, Max 20x), a user enables credits under Settings > Usage on the Claude web app, adding a payment method and a spending limit; someone who subscribed through a mobile app still sets it up on the web. Team and Enterprise admins enable it for the organization. Turning on credits doesn't change the session limit, which still resets every five hours.
Enterprise depends on seat type. Per Anthropic's footnote, standard seats have no included Fable 5 allowance and need credits to use it at all, while premium seats include Fable 5 through July 7 and require credits after that.
The safeguards are tighter
The redeployed Fable 5 ships with a stronger safety classifier. Anthropic said the June 12 halt traced to a safeguard-bypass technique found by Amazon researchers, and that the new classifier blocks that technique in more than 99% of cases — its own figure. The tradeoff, the company added, is that ordinary coding and debugging requests may get flagged more often than before.
When a request is blocked, the user is notified and it is routed to Opus 4.8 instead. In the API, a Fable 5 refusal is not an error: it comes back as an HTTP 200 response with stop_reason set to "refusal".
The clouds and Mythos 5 aren't back yet
Anthropic said it would re-enable Fable 5 on AWS, Google Cloud, and Microsoft Foundry "as quickly as possible." At the time of the statement, all three were still pending.
Mythos 5 is narrower. It is the same underlying model as Fable 5 with the safeguards stripped out, and Anthropic said it restored access only for a set of U.S. organizations, following U.S. government approval on June 26. Widening it to the broader group of domestic and international Glasswing partners is still being worked out with the government. Whether Korean companies can use Mythos 5 again is not confirmed in this statement.
Sources
Primary sources
- Anthropic: Redeploying Fable 5
- Anthropic: Statement on the US government directive to suspend access to Fable 5 and Mythos 5
- Claude Platform Docs: Introducing Claude Fable 5 and Claude Mythos 5
- Claude Help Center: Manage usage credits for paid Claude plans
- Claude Help Center: Manage usage credits for Team and seat-based Enterprise plans
- Anthropic: Supported countries & regions
Companion pieces
- Stopped by a government order: Claude Fable 5 and Mythos 5
- The same Claude model, but not for everyone: Mythos 5 returns for approved U.S. organizations only