Free Access Ends, a "Half Allowance" Makes It Permanent — The Day Claude Fable 5 Found a Home in Max/Team Premium, While Pro Got Shifted to Credits

On July 20, Anthropic makes its top-tier Fable 5 model a permanent fixture in Max and Team Premium — but limits it to 50% of the reduced usage cap, while Pro and Team Standard get a one-time $100 in credits before falling back to metered billing. Here's what it means for pricing and how you use it.

Share
Free Access Ends, a "Half Allowance" Makes It Permanent — The Day Claude Fable 5 Found a Home in Max/Team Premium, While Pro Got Shifted to Credits

On July 18, Anthropic announced a rethink of how it delivers its top-tier model, "Claude Fable 5." Starting July 20, Fable 5 is permanently included in both the Max and Team Premium plans. For Pro and Team Standard, however, it drops out of the plan's standard allowance and switches to a pay-as-you-go credit model. The move draws a line under the repeatedly "extended free access" of early July, settling the matter by concentrating the model in the higher-tier plans.

A "half allowance" in exchange for a permanent slot

This change treats plans in distinctly different ways. Max and Team Premium gain access to Fable 5, but not on the same footing as the regular allowance. First, the overall usage limit for the higher-tier plans is cut by roughly 33% (the end of the bonus increases that had been rolled over repeatedly), and Fable 5 is capped at 50% of that reduced limit. For Pro and Team Standard, standard-allowance access effectively ends, replaced by a one-time $100 worth of credits. Once those are used up, usage falls under the same pay-as-you-go rates as the API.

PlanMonthlyHow Fable 5 is handled from 7/20
Max 5x$100Included in plan (up to 50% of the reduced limit)
Max 20x$200Included in plan (up to 50% of the reduced limit)
Team PremiumIncluded in plan (up to 50% of the reduced limit)
Pro$20Pay-as-you-go credits only (one-time $100 → API rates thereafter)
Team StandardPay-as-you-go credits only (same as above)

Anthropic frames this "half allowance" as a line it is drawing: it will not hand Fable out to every plan at the same usage levels as Sonnet, Opus, and Haiku. The aim is to deliberately differentiate a frontier-class model by keeping it to a tighter allowance even within the higher-tier plans.

The "metered wallet" now handed to Pro developers

The biggest impact falls on Pro ($20/month) and Team Standard users. Fable 5, which they could touch within their plan during the free-access period, becomes a credit-consuming resource starting July 20. Once the $100 in credits is spent, everything beyond that rides on the same pay-as-you-go billing as a production app. For reference, Fable 5's API pricing is unchanged, with rough figures as follows.

CategoryRate (per 1M tokens)
Input$10
Cached input$1
Output$50

Running a top-tier model in a coding agent tends to inflate token counts, as codebases and logs pile up on the input side. For the Pro tier, this is a shift from "try it for a flat fee" to "pay for what you use," and if you rely on it regularly, it's worth noting that monthly costs become harder to predict.

Why tighten the limits now

Behind the move lie both supply-and-demand and competitive factors. Anthropic acknowledged that "demand for Fable has been hard to manage and stressful for users too," while saying it will "continue to invest in capacity." It's reported that the company initially considered dropping Fable entirely from the subscription tiers before reversing course. The Decoder points to contributing factors including the arrival of OpenAI's GPT-5.6 Sol at roughly a third of Fable 5's cost, and price pressure from Chinese-made models. Deciding "which plans get the top-tier model, and how much" has now become part of the price competition itself.

The questions that remain behind the convenience

With this reshuffle, higher-tier users can expect stable access to Fable 5. At the same time, you could say access to a frontier-class model has been sharply stratified by how much you pay. The Decoder describes the change as one that "pushes Pro users toward API pricing." It's not a definitive judgment, but the concern can be raised that the more a lower-tier developer uses the top model day to day, the harder metered costs become to predict. If you're weighing adoption, the practical approach is to first estimate — from your expected monthly token volume — whether your usage fits within "the higher-tier plans' half allowance," or whether you should plan around metered billing with credits factored in.

References: Anthropic to add Claude's Fable 5 model to Max, Team Premium plans at 50pc of usage limits (DAWN) / Anthropic slashes Claude Fable 5 limits in Max and Team Premium (The Decoder) / Claude Fable 5 Comes to Max and Team Premium: Pricing Impact (AI Pricing Guru) / Claude Fable 5 promotion extended after backlash over early cutoff (Android Authority)

Read more

Making It Wait for "Jobs That Run Over an Hour": Codex 0.152 Adds Ceiling Dials for MCP Output Volume and Execution Time, and Turns the Planning Tool Off by Default

Making It Wait for "Jobs That Run Over an Hour": Codex 0.152 Adds Ceiling Dials for MCP Output Volume and Execution Time, and Turns the Planning Tool Off by Default

Codex v0.152.0 on August 31 and its next-day fix release added explicit ceilings on MCP tool output volume and execution time, and switched the planning tool off by default. Here's a rundown of the changes that matter for long-running unattended and semi-autonomous agent operation.

By FF
The CLI's Default Model Just Swapped In a Million-Token Brain — Claude Code v2.1.257 Makes Fable 5.1 the Standard and Adds a 'Containment Escape' Checkpoint to Auto Mode

The CLI's Default Model Just Swapped In a Million-Token Brain — Claude Code v2.1.257 Makes Fable 5.1 the Standard and Adds a 'Containment Escape' Checkpoint to Auto Mode

Claude Code v2.1.257, released September 1, 2026, swaps its default model to Fable 5.1 with its one-million-token context. It also adds guardrails to auto mode that stop credential retrieval and out-of-scope reads from slipping through. Here's a rundown of the changes that matter to developers.

By FF
"This Is an Authorized Exercise"—How the Aurora Ransomware Gang Insisted, While Making Cursor's AI Agent Do the Actual Intrusion Work

"This Is an Authorized Exercise"—How the Aurora Ransomware Gang Insisted, While Making Cursor's AI Agent Do the Actual Intrusion Work

Gambit Security and CloudSEK report that the ransomware group Aurora abused Cursor's AI agent for real intrusion work. Posing the tasks as an "authorized exercise" to slip past the safeguards, they had it handle reconnaissance and privilege takeover on the back of stolen credentials—a warning that a

By FF
One in Three Companies Now Choose to Build Rather Than Buy — McKinsey Measures How Coding Agents Are Reshaping the Procurement Decision

One in Three Companies Now Choose to Build Rather Than Buy — McKinsey Measures How Coding Agents Are Reshaping the Procurement Decision

McKinsey's annual survey found that about 30% of respondents passed on buying software because they could build it in-house with coding agents. We unpack the procurement shift from buying to building — and the current reality that productivity is up while profits stay flat.

By FF