Anthropic is giving Claude Fable 5 a longer runway inside paid subscriptions. The company has extended promotional, subscription-included access to its Mythos-class flagship through July 19, and the 50% bump to Claude Code’s weekly rate limits is riding along with it on the same deadline.
The mechanics of the extension are straightforward. Pro, Max, Team, and premium seats on seat-based Enterprise plans can keep using Fable 5 for up to 50% of their weekly usage limits at no extra cost, through July 19 at 11:59:59 PM PT. Once that allowance runs out, users can either switch billing over to usage credits and keep going on Fable 5, or fall back to another Claude model and stay within their normal weekly caps. Nothing needs to be claimed or toggled on — it’s already live across Claude on the web, Claude Mobile, Claude Desktop, Claude Cowork, and Claude Code (version 2.1.170 or later required there). Standard seats on legacy Enterprise plans and API usage remain outside the promotion entirely; API access to Fable 5 has always been billed separately at standard rates.
This is actually the second time Anthropic has pushed this deadline back. Fable 5 only returned to global availability on July 1, after a Commerce Department export-control order forced Anthropic to pull both Fable 5 and Mythos 5 offline worldwide for nineteen days starting June 12. When access came back, the original plan was for the 50% subscription allowance to run out on July 7, after which Fable 5 would move to paid usage credits at $10 per million input tokens and $50 per million output tokens — among the priciest rates Anthropic has ever listed for a public model. That July 7 cutoff came and went without much fanfare, but this new extension to July 19 is the first time Anthropic has put a fresh date on the record.
The timing lines up almost exactly with a rough couple of weeks for Anthropic on the competitive scoreboard. OpenAI’s GPT-5.6 family — Sol, Terra, and Luna — went from limited preview to general availability on July 9, and Sol in particular has been landing close to Fable 5 on several fronts. On the Artificial Analysis Intelligence Index, Sol at max reasoning effort trails Fable 5 by a single point — 59 to 60 — while costing roughly a third as much per task. On the Coding Agent Index, Sol actually pulls ahead of Fable 5 outright, running about 40% cheaper per task inside Codex than Fable 5 does inside Claude Code. Sol’s Terminal-Bench 2.1 score of 91.9% in ultra mode also edges past Claude Mythos 5’s 88.0%, a benchmark OpenAI leaned on hard in its own launch messaging.
Anthropic hasn’t tied the extension to any of this directly. But the sequence of events reads that way. Days after OpenAI’s GPT-5.6 livestream wrapped, Anthropic quietly reset rate limits for every user, five-hour and weekly caps included, prompting OpenAI’s Tibo Sottiaux to needle the company on X with a three-word “I smell fear” that racked up hundreds of thousands of views. Keeping Fable 5 inside the subscription tier for another week, alongside heavier Claude Code limits, is a similar kind of move: it keeps developers and power users testing Fable 5 against Sol on Anthropic’s dime rather than nudging them toward usage credits or, worse, toward a competitor’s cheaper flagship.
There’s also a simpler explanation sitting underneath the competitive one. Fable 5 spent nearly three weeks completely dark this summer because of the export-control order, and Anthropic has spent the time since trying to rebuild goodwill with users who had workflows built around a model that vanished overnight without warning. A longer free window gives that goodwill more time to compound before the usage-credit meter starts running at $10/$50 per million tokens. Whether the driving force is Sol’s benchmark numbers, the aftertaste of the suspension, or both, the practical outcome for anyone on a paid Claude plan is the same: more Fable 5, at the regular subscription price, for an extra week.