Fable 5 is back, and it's live in Claude Code

Hook

The most capable model Anthropic has ever put on the open market went dark on June 12, blocked by a US export control. As of July 1 it's back on, including inside Claude Code, where operators actually work.

The news

Anthropic turned Claude Fable 5 back on July 1, after the US government lifted the export controls it applied on June 12. The block came after Amazon researchers found a way around the model's safeguards, "prompting it so that it identified a number of software vulnerabilities," and the directive forced Anthropic to cut off every user, since it could not verify nationality in real time. Fable is restored to the Claude Platform, Claude.ai, Claude Code, and Claude Cowork, with AWS, Google Cloud, and Microsoft Foundry to follow. It ships with a hardened classifier Anthropic says blocks the reported technique "in over 99% of cases," and redirects flagged requests to Opus 4.8. Anthropic's post is here.

On your desk

If you run on a harness, this is an engine upgrade you don't have to install: Fable is state-of-the-art on nearly every benchmark, and it's now selectable in Claude Code, so your existing workflows point at a stronger model the second you switch. One catch worth stating plainly. On Pro, Max, Team, and select Enterprise plans, Fable is included up to 50% of your weekly limits only through July 7. After that you pay usage credits, at $10 per million input tokens and $50 per million output.

What you should do next

Open Claude Code, switch the model to Fable 5, and run something hard before July 7 while it's still on the house.