Anthropic has extended complimentary access to Claude Fable 5, its most capable large language model (LLM), for paid subscribers through July 19, 2026 — the third such extension since the model was redeployed on July 1 following a nearly three-week government-ordered suspension.
Background: A Turbulent Launch
<cite index="17-10">Anthropic released Claude Fable 5 and its companion model Claude Mythos 5 on June 9, 2026, the most capable models in the company's history, noted for their advanced reasoning and code analysis capabilities.</cite> The debut was short-lived. <cite index="17-1,17-2">On June 12–13, U.S. Commerce Secretary Howard Lutnick sent Anthropic CEO Dario Amodei a letter directing the company to place both models under export controls, prohibiting access by any foreign national whether inside or outside the United States.</cite> <cite index="22-18">Because a user's nationality cannot be verified in real time at the API layer, Anthropic had only one option: switching off both models for all customers worldwide.</cite>
<cite index="18-4,18-5">On June 30, the U.S. Commerce Department lifted the export controls it had imposed on Fable 5 and Mythos 5, and Fable 5 returned to users on July 1 across Claude.ai, the Claude Platform, Claude Code, and Claude Cowork.</cite> Upon redeployment, <cite index="19-4">the model was included for up to 50% of weekly usage limits through July 7 for users of the Pro, Max, Team, and selected enterprise plans.</cite>
Three Extensions in Two Weeks
<cite index="1-4,1-5,1-6">Previously, Anthropic said Fable 5 would be included in Pro, Max, Team, and premium Enterprise subscriptions only through July 7; after that, usage credits would be required. Anthropic then extended Fable 5 usage via paid plans until July 12.</cite> The company has now extended the window a third time. <cite index="1-8">"We've extended this promotion through July 19, 2026 at 11:59:59 PM PT," Anthropic noted in a support document.</cite> <cite index="1-9">"The 50% increase to Claude Code weekly usage limits has also been extended through the same date."</cite>
<cite index="1-10,1-11,1-12">Subscribers can use Fable 5 for up to 50% of their weekly subscription limits at no extra cost, with Fable 5 drawing from the same weekly usage pool as other Claude models — though Anthropic warns it consumes that limit faster than other models.</cite> <cite index="2-10">Once users exhaust 50% of their weekly limit or after the free access period ends, they must purchase prepaid usage credits priced at $10 per million input tokens and $50 per million output tokens.</cite>
<cite index="2-11">Anthropic has stated that its goal is to transition Fable 5 into a formal subscription product as soon as sufficient computing capacity is secured.</cite> The repeated extensions suggest that scaling infrastructure to support the model at subscription-tier volume remains an ongoing challenge.
Competitive Context: GPT-5.6 Enters the Field
The extensions coincide with a significant competitive development. <cite index="10-1">GPT-5.6, a new LLM developed by OpenAI, was released publicly on July 9, 2026.</cite> <cite index="12-7">GPT-5.6 comes in three variants: Sol (its workhorse), Terra (a more intermediate option), and Luna (its budget-friendly option).</cite>
OpenAI has positioned the release as a direct challenge to Anthropic. <cite index="12-10,12-11">OpenAI calls Sol its "best coding model yet" and has explicitly compared it to Fable 5, claiming Sol "sets a new state of the art at 80, 2.8 points above Fable 5, while using less than half the output tokens, taking less than half the time, and costing about one-third less."</cite> <cite index="9-26">GPT-5.6 is priced per million tokens: Sol at $5 input / $30 output, Terra at $2.50 input / $15 output, and Luna at $1 input / $6 output</cite> — making its top tier significantly cheaper per token than Fable 5 at API list prices.
<cite index="8-8,8-9,8-10">Anthropic's moves align with a broader industry trend: as flagship AI models become more expensive to run, companies including Anthropic, OpenAI, and Google are experimenting with hybrid pricing rather than unlimited subscriptions, letting users experience the newest models first, then charging only those who rely on them heavily.</cite>
Anthropichas not publicly attributed the access extensions to GPT-5.6's release; its official communications have cited capacity planning and the disruption caused by the export-control episode. Nonetheless, the timing places Fable 5's prolonged promotional window squarely alongside OpenAI's most aggressive comparative marketing push to date, underscoring how frontier model competition is increasingly shaping near-term user-facing product decisions.