Correction: Claude Fable 5 Is on Max Plans at 50% of Weekly Limits

Verdict: we were wrong about access, and the rating still holds. Our directory entry for Claude Fable 5 said the model was credit-only with no subscription path. Anthropic's own plan table lists Fable at 50% of weekly limits on Max 5x, Max 20x and Team premium seats. Fable 5 stays conditional, on two constraints that have nothing to do with access: the $10/$50 per 1M API price and the safety-classifier fallback. Claude Fable 5.1 sits at the same base price with cheaper cache hits, and for routine work our directory points to Claude Opus 5 if you already run it and Claude Opus 5.5 for new workloads.

What happened

We re-checked Claude Fable 5 against Anthropic's published plan pricing on 14 September 2026 and found our own entry wrong. We checked the page again on 12 October 2026. The plan table reads:

Our verdict summary said Fable 5 had "credit-only pricing at $10/$50 per 1M tokens with no subscription path". The first half is right. The second half was not: Max subscribers and Team premium seats have a subscription path. The entry now reads "Subscribers reach Fable through Claude Max at 50% of weekly limits, or through usage credits on Pro."

Two caveats we are not going to paper over. First, the table gives the percentage, not the mechanic. "50% of weekly limits" carries only a footnote reading "Usage limits apply." It does not say whether Fable draws down your weekly allowance at double rate or is capped at half of it, and the page does not publish the underlying weekly allowance. Second, the table lists a single "Fable" row. It does not break out Fable 5 from Fable 5.1, so read the row as covering the Fable line rather than one version.

Why it matters

There are two costs here, and we caused one of them.

Max starts at $100 per month. Pro is $17 per month on annual billing, or $20 billed monthly. A team reading our entry and concluding it had to buy API credits at $10 per 1M input and $50 per 1M output to touch Fable 5 would have opened a second billing relationship for capability it may already have been paying for. At the volumes where a frontier model earns its keep, that is a real number.

The second cost is ours, and it is not that we missed the news. We reported it: Fable 5 gets subscription access covered Max and Team Premium inclusion at 50% of plan limits on July 20. The story reached the post and the model's pricing note. It never reached the verdict summary, which still said "no subscription path" on 14 September, eight weeks after we published the opposite. That is a process defect: a fact we published that never propagated to the directory entry readers actually consult. It is why this is a post rather than a silent edit.

What changes for you

• On Max 5x or Max 20x: Fable is inside your plan at 50% of weekly limits. Try the plan path before you provision API credits.
• On Team: the table lists Fable at 50% of weekly limits on premium seats. Check which seat type you hold before you plan around it.
• On Enterprise: the table lists Fable as included ("Yes").
• On Pro: nothing changes. Anthropic lists Fable as usage credits, so budget it as pay-per-use on top of the subscription.
• On Free: no Fable access.
• On the API: the rate card is untouched. Fable 5 is $10 per 1M input, $50 per 1M output, cache hits at $1 per 1M, and 5-minute cache writes at $12.50 per 1M. The Batch API halves input and output to $5 and $25.
• If cache reads dominate your bill: Claude Fable 5.1 is the same $10/$50 base with cache hits at $0.25 per 1M instead of $1, a fourfold difference on that line.

The conditional verdict on Fable 5 holds on its two remaining constraints. The first is price: $10/$50 per 1M is double Claude Opus 5's $5/$25. The second is the safety classifiers. When they flag a request as cybersecurity, biology and chemistry, or distillation, Fable hands the response to a less capable model (Claude Opus 4.8 at launch; Anthropic's August 7 safeguards update names Claude Opus 5 for biology), and Anthropic says users are told when that happens. Anthropic puts the share of Fable sessions with no fallback at more than 95%, and its own updates concede false positives: about 85% fewer biology fallbacks after the August update, and 60% fewer cybersecurity false positives with Fable 5.1. For routine work, Claude Opus 5 at $5/$25 per 1M stays the predictable pick for teams already on it, and new workloads should start on Claude Opus 5.5 at $4/$20 per 1M.

FAQ

Is Claude Fable 5 included in the Claude Max plan? Yes. Anthropic's plan table lists Fable at 50% of weekly limits on both Max 5x and Max 20x. It is not credit-only.

Is Fable included on Team plans? On premium seats. The table lists Team at 50% of weekly limits on premium seats, so a standard seat does not carry it.

Didn't you already report this? Yes. Fable 5 gets subscription access covered Max and Team Premium inclusion at 50% of plan limits on July 20. That reporting was right. The directory entry it should have updated was not, and closing that gap is what this post is for.

Does Claude Pro include Fable 5? Not as plan usage. Anthropic lists Fable on Pro as "usage credits", so it bills as pay-per-use rather than drawing on the plan.

What does "50% of weekly limits" actually mean? Anthropic publishes the phrase without defining the mechanic and without publishing the underlying weekly allowance. We are not going to guess. Read it as: Fable is available on Max, at a reduced share of whatever your weekly ceiling is.

Did the Fable 5 API price change? No. $10 per 1M input and $50 per 1M output, unchanged. Only our description of the subscription path was wrong.

Does this change your verdict? No. The rating stays conditional on two constraints: the $10/$50 per 1M API price and the safety-classifier fallback on flagged cybersecurity, biology, chemistry and distillation requests. Access was never a valid reason.