Fable 5.1 Is In Your Max Plan. It Is Not Free.
Fable is included on Max at up to half your weekly limits. Input and output cost twice Opus 5; cache reads cost half. Whether it is cheaper per accepted task depends on workload and an efficiency claim that needs testing.
If you pay $200 a month for Claude Max 20x, Fable has been part of your subscription since 20 July 2026. Up to half of your weekly usage limits may go to Fable at no extra charge. Read that precisely: it is a ceiling inside the allowance you already have, not a second pool bolted on top. A lot of people who wrote Fable off as the expensive model never noticed it moved at all.
Fable 5.1 and Mythos 5.1 shipped today, 1 September 2026, nearly three months after Fable 5. The sticker price did not move: $10 per million input tokens, $50 per million output — exactly twice Opus 5 on both. What did move is cache reads, cut 75% from $1.00 to $0.25 per million, which is half what Opus 5 charges for the same thing. So Fable is twice the price on fresh tokens and half the price on cached ones.
So the interesting question is not whether Fable is expensive. It is whether the price you actually pay, in dollars or in weekly allowance, is anywhere near double.
Fable 5.1 is included in Max at no extra cost, capped at 50% of your weekly limits. Past that you switch back to Opus or wait for the reset; continuing bills usage credits at API rates.
Input and output cost exactly 2x Opus 5. Cache reads cost half. Agentic coding is cache-read heavy, so the effective gap is much smaller than the headline.
On the illustrative session below the premium is about 25%, not 100%. Fable reaches parity at 80% of Opus's tokens if all three traffic categories scale together.
One Anthropic partner reports half the tokens at twice the speed. No independent controlled comparison is cited here, and you can settle it on your own repos in a week.
Headless work draws from the subscription when authenticated through an eligible subscription flow; API keys use API billing. The proposed June 15 billing change was paused; check the current policy before budgeting.
Anthropic describes Mythos 5.1 as the same underlying model with different safeguards and access through trusted programs. Vetted cyberdefence and life-sciences organisations only.
What the 50% cap actually means
Anthropic's support documentation now covers Fable 5 and Fable 5.1 explicitly. On Max and Team Premium you may spend up to 50% of your weekly usage limits on Fable models at no extra cost. The same document is blunt about which pool that is: Fable models draw from your plan’s regular weekly usage limits and use them faster than other Claude models, and you can never use more than your weekly limit. There is one bucket. Exhaust it and switching to Fable rescues nothing, because Fable was drinking from the same bucket the whole time. Hit that ceiling and you switch model or wait for the weekly reset. Continuing on Fable is billed to usage credits at API rates, and only if you have enabled them.
Fable access and promotional credits depend on the plan and model version. The current plan documentation says Pro and Team Standard use paid credits for Fable, and no equivalent Fable 5.1 promotional credit is offered. Verify the current terms rather than assuming an earlier offer applies.
As checked 20 September 2026, Anthropic’s plan documentation requires Claude Code 2.1.170 or later for Fable 5 and 2.1.255 or later for Fable 5.1. The model ID is claude-fable-5-1.
Subscription allowance is not a dollar or token budget. The plan documentation says Fable consumes the regular allowance faster, but that does not establish an exact 2× weighting. API prices cannot supply that conversion. Measure allowance consumed per accepted task separately from API cost.
The arithmetic, out loud
Fable 5.1 cache writes cost $12.50 per million tokens for five minutes and $20 for one hour. Its standard input/output rates match Opus 5 fast mode, while its $0.25 cache reads are one quarter of Opus fast’s $1.00. Standard Opus batch input/output rates remain half of Fable batch’s; compare the complete token mix and applicable modifiers.
The September 2026 announcement attributes lower Fable 5.1 token-billed costs to reduced cache-read pricing, relative to Fable 5. That does not establish savings against Opus. The price table above uses the official API rate card, checked 20 September 2026. Rates and service modifiers should be checked before budgeting.
A worked bill for one long agent session
Assume a session that burns 2M ordinary input tokens, 400K output tokens and 40M cache reads — an agent repeatedly rereading a large cached repository. These are illustrative numbers, not measurements, and cache-creation charges are excluded.
On this illustrative mix, at identical token volumes, Fable costs 25% more rather than 100% more. The cache line does all the work: it is the only row where Fable is cheaper, and in agentic coding it is usually the biggest row. This proves one traffic mix, not a universal premium, and it excludes cache-creation charges.
The break-even is worth stating as algebra, because it lets you compute your own. With I million input, O million output and C million cache reads at equal volumes, the two bills are equal when C = 20I + 100O. For this example that is 80M cache reads. We assumed 40M, so Fable stays more expensive on volume alone — until token efficiency enters.
And that threshold is low. If input, output and cache reads all fall in the same proportion, Fable reaches parity at 80% of Opus's volume. Halve all three and the bill lands at $25, 37.5% under Opus. Real agents will not reduce every category proportionally, so treat 80% as the shape of the answer rather than the number.
A partner claim to test
Dan Shipper at Every reports Fable 5.1 running roughly twice as fast as Opus 5 while using about half the tokens. If that reproduces on your work at comparable acceptance quality and retry rates, the modelled bill favours Fable: double the per-token price times half the tokens is a wash, and the cheaper cache reads then push you below Opus.
Dan Shipper’s claim is an early-access partner report quoted in Anthropic’s launch announcement, not an independent controlled measurement.
Public anecdotes about verbosity and speed are not controlled cost comparisons. This article does not use them to establish which model is cheaper per accepted task. The useful test holds task difficulty and acceptance criteria constant and counts retries and corrections.
A model can produce longer prose while finishing a task with fewer retries. API cost depends on billed token categories and other applicable charges, not the number of turns alone. Subscription allowance consumption is a separate measurement.
The benchmarks are more mixed than the headlines
The official launch table, checked 20 September 2026, gives Opus 5 a Terminal-Bench 4.0 score of 52.3%; 42.0% is Fable 5’s score. The corrected table retains three explicitly reported evaluations. Scores depend on effort, harness and safeguard handling; they do not establish universal task performance.
Headless billing depends on authentication
Anthropic proposed moving Agent SDK usage to a separate credit arrangement, then paused the change in its June 15 update. The old proposal remains on that page as historical text and is not the current billing rule.
When authenticated through an eligible subscription flow, Agent SDK and claude -p usage draw from that subscription’s limits. API-key usage remains pay-as-you-go. Headless execution alone does not choose the billing route, and a third-party client must also comply with the provider’s authentication rules.
For anyone running unattended agents: subscription-authenticated headless runs consume the same allowance you need for the next morning's interactive work, and Fable's 50% cap applies to that shared pool rather than getting its own headless budget. If you sized your automation around the announced credit system, resize it. Related: Your Agents Eat Your Subscription.
Run the one-week test instead of guessing
Turn Fable 5.1 on and keep standard Opus 5 as the control — not fast mode, which shares Fable’s input/output rates but has different cache-read pricing. Alternate comparable tasks between them for a week rather than sending all the hard work to one model. Track this:
Mythos 5.1 has a separate access policy
Anthropic describes Fable 5.1 and Mythos 5.1 as the same underlying model with different safeguards. Mythos access is limited to trusted programs for eligible cybersecurity and life-sciences work; this is distinct from ordinary Max-plan inclusion. See the official announcement for the current program scope.
Bottom line
Eligible Max plans include a share of Fable use within their regular limits. Separately, the stated illustrative API token mix yields a 25% premium over standard Opus. That arithmetic does not determine subscription allowance weighting or an individual plan’s value.
No independent controlled comparison is cited here. The partner efficiency report is a reason to test on representative tasks, not to assume a particular saving.
Compare accepted tasks, correction effort, elapsed time and allowance consumption on the subscription route. For the illustrative API mix only, a proportional 20% reduction across the stated token categories reaches break-even; it is not a subscription threshold. Keep paid extra usage settings explicit while testing.
The 50% is a ceiling, not a gift. Make Fable earn the share of the allowance it takes.
