Claude Fable 5 Price: Is It Free, Usage Credits, and Access

Short answer: Claude Fable 5 was free on Claude Pro, Max, Team, and seat-based Enterprise plans only from June 9 through June 22, 2026. On June 23, Anthropic removed it from those plan limits, and continued use now requires usage credits billed at API rates. On the Claude API directly, Fable 5 runs $10 per million input tokens and $50 per million output tokens, double the price of Opus 4.8 and the most expensive model Anthropic has priced for general use.

There is one more thing you need to know before you plan around any of this: as of June 12, 2026, Fable 5 is temporarily suspended for everyone under a US government export-control directive (covered below). This guide breaks down exactly what Fable 5 costs, how the usage-credits switch works mechanically, what the free window was, and what the current access situation means for your spend.

Is Claude Fable 5 free?

Claude Fable 5 was free, but only for thirteen days. From June 9 through June 22, 2026, it was included at no extra cost on Pro, Max, Team, and seat-based Enterprise plans. On June 23 that window closed: Anthropic removed Fable 5 from those plan limits, and using it now draws from prepaid usage credits billed at standard API rates. On the API it was never free, it is paid from day one at $10/$50 per million tokens. The claude.ai free tier does not include Fable 5 at all.

So if someone asks "is Fable 5 free," the honest answer is: it was a free trial inside paid subscriptions for two weeks, and that trial is over. Below is exactly what it costs now and how the credits system works. (And see the access note at the end, because right now nobody can reach it regardless of how they pay.)

Claude Fable 5 API Pricing

Here is the full API rate card for Fable 5, set against Opus 4.8 and Mythos Preview so you can see where it lands:

Three things matter here. Fable 5 is priced at exactly double Opus 4.8 on both input and output. It is less than half the price of Mythos Preview, the frontier-class tier that stays restricted to Project Glasswing partners. And the 90% prompt-caching discount carries forward: a cache read costs $1 per million tokens, one tenth of the standard input rate, so heavy-context agentic workloads that reuse the same system prompt or codebase across calls pay far less than the headline number suggests.

The token math has one large upside worth stating plainly: Fable 5 ships the full 1M-token context window at standard pricing. A 900K-token request bills at the same per-token rate as a 9K-token request, with no long-context surcharge. For the long-horizon, whole-codebase work Fable 5 is built for, that flat rate is the difference between "expensive" and "unworkable."

Two pricing modifiers stack on top. Batch processing cuts both input and output in half ($5/$25 per million) for non-urgent jobs. And US-only inference, set with inference_geo: "us", applies a 1.1x multiplier across every token category if you need guaranteed US data routing. Global routing is the default and uses standard pricing. The model ID is claude-fable-5, and it shipped on the Claude API, Claude Code, consumption-based Enterprise, plus AWS, Google Cloud's Vertex AI, and Microsoft Foundry from day one. For where Fable 5 lands on raw capability versus its price, see the Fable 5 benchmarks and specs.

What Changed on June 23

For the first two weeks, Fable 5 behaved like every other model on your subscription. You selected it, you used it, and it drew down your normal plan usage. The wrinkle was the weighting: inside subscription plans, Fable 5 counted roughly double the usage of Opus toward your limits. So even during the free window, a Fable 5 session burned through your plan allowance about twice as fast as the same work on Opus 4.8.

On June 23, the arrangement ended. Anthropic removed Fable 5 from the included plan limits entirely. Since that date, the model still appears in your picker, but using it draws from usage credits rather than your bundled allowance. No credits, no Fable 5 beyond the free window.

Anthropic has been explicit that this is a capacity decision, not a permanent pricing tier. Its stated intent is to "restore Fable 5 as a standard part of subscription plans" once capacity allows, and to extend the included window if capacity holds. That is a real commitment, but it is also an open-ended one. There is no announced date for restoration, so the practical assumption to plan around is: a Fable 5 subscription now costs money on top of your monthly fee.

How Fable 5 Usage Credits Work

Usage credits are Anthropic's prepaid overage system for paid plans. They are available on Pro, Max 5x, and Max 20x, and the same mechanism extends to Team and seat-based Enterprise. Here is the mechanic, straight from Anthropic's usage-credits documentation:

They activate only after you exhaust your included plan usage. Credits are not a substitute for your subscription. You hit your normal limit, get a notification, and then, if credits are enabled and funded, you keep going. Your standard plan usage still resets every five hours as usual; credits only cover what spills over that ceiling.

You prepay them, in dollars. Go to Settings > Usage, enable the feature, click Add funds, and enter an amount. There is a daily redemption limit of $2,000. You can also set up auto-reload so funds top up automatically when your balance drops below a threshold you choose.

They are billed at standard API rates. This is the key point for Fable 5. Credits are not consumed at some separate consumer rate; they draw down at the same per-token API prices in the table above. Every token counts: chat messages, Claude Code terminal usage, Research mode sessions, and project file content all consume credits once you are past your plan limit.

When credits run out, you revert to your plan's included usage. If the balance hits zero or the feature is disabled, you simply lose overage capacity until you add more funds. Nothing breaks; you just can't exceed your plan allotment.

One operational note: if you bought your subscription through a mobile app store, you can only enable and purchase credits on the web version of Claude, not in the mobile app.

What Subscribers Should Do Now

The free window is over and access is paused, so the playbook is about budgeting for the model you'll reach for when it returns, not racing a deadline.

Budget for credits if you plan to keep using it. Since June 23, continued Fable 5 use on a subscription is prepaid API spend. Enable usage credits, set an auto-reload threshold you're comfortable with, and remember the daily $2,000 ceiling. If you don't fund credits, your access stops at your plan's included allowance. One small mercy in the billing design: when a request gets rerouted to Opus 4.8 by the safeguard classifiers, you are not charged Fable 5 prices for it, so the safety fallback never silently inflates your bill.

Control your effort level, not just your model. Most of a Fable 5 bill is output tokens at $50 per million, and output volume tracks how hard you let the model think. Capping reasoning effort on routine work, the same discipline covered in our Ultracode effort guide, is the single biggest lever on credit burn once you are paying per token.

For heavy use, compare the API directly against subscription credits. Since credits bill at standard API rates anyway, the choice between "subscription plus credits" and "a straight API key" comes down to whether your base plan usage still pulls its weight. If you are blowing past plan limits constantly, an API key with prompt caching and batch discounts may be the cleaner accounting. If your overage is occasional, credits on top of your existing plan are simpler. Either way, the higher Claude Code rate limits that shipped alongside recent releases give you more headroom before credits ever kick in.

When the 2x Price Is Worth It

Doubling your token cost is not a default decision. Fable 5 earns its price on a specific class of work: long-horizon, complex tasks where it leads the field and where a single better outcome justifies the spend. Think multi-hour autonomous agent runs, intricate codebase migrations, or analysis where a wrong answer is expensive to catch later.

For most day-to-day coding, Opus 4.8 at half the price is the rational default. It is Anthropic's strongest generally available model on cost-to-quality, completes full agentic tasks end-to-end, and is the model most teams should reach for first. The honest framing: Fable 5 is the model you escalate to, not the model you run all day. Reserve it for the work where it measurably wins, and let Opus 4.8 carry the volume. For a fuller breakdown of which model fits which task, see our model selection guide.

If your work spans research, content, and validation rather than pure development, the cost calculus is the same: match the model to the task. ClaudeFast's Growth Kit is built around that discipline, with workflows that route work to the right tier instead of defaulting to the most expensive option for everything.

Current Access: The June 12 Suspension

There is a wrinkle that overrides everything above. On June 12, 2026, the US government issued an export-control directive suspending access to Fable 5 and Mythos 5 for any foreign national, inside or outside the United States, including Anthropic's own foreign-national employees. The only way Anthropic could comply was to disable both models for every customer. So as of now, the usage-credits arrangement is moot in practice: nobody can run Fable 5, on a subscription or on the API, until access is restored.

Two things keep this from being a reason to write off the model. First, the directive is narrow. It traces to a non-universal jailbreak (essentially asking the model to read a codebase and patch software flaws), and Anthropic has called the suspension "a misunderstanding" that it is "working to restore access as soon as possible." Second, every other Anthropic model is unaffected, so Opus 4.8 remains available at $5/$25 and stays the right default for the vast majority of work regardless of what happens with Fable 5.

The practical read: keep your credit budgeting ready for when Fable 5 returns, but do not block any current workflow on it. There is no confirmed restoration date. We will update this guide when access changes.

Frequently Asked Questions

Is Fable 5 free on Claude Pro? It was, but only from June 9 through June 22, 2026. During that window it was included at no extra cost on Pro, Max, Team, and seat-based Enterprise. On June 23 it was removed from plan limits, and continued use now requires usage credits. It was never free on the API.

What happened on June 22? June 22 was the last day Fable 5 was included free on subscription plans. On June 23, Anthropic removed it from plan limits, so using Fable 5 on a subscription now draws from prepaid usage credits billed at API rates. Anthropic intends to restore it as a standard plan feature once capacity allows, but has not announced a date.

Can I use Fable 5 right now? Not at the moment. A US government export-control directive on June 12, 2026 forced Anthropic to disable Fable 5 and Mythos 5 for all customers. Every other Anthropic model, including Opus 4.8, is unaffected. Anthropic is working to restore access and has not given a timeline.

How much does Fable 5 cost? On the Claude API, $10 per million input tokens and $50 per million output tokens, with a 90% prompt-caching discount on cached reads ($1 per million) and the full 1M-token context window at standard pricing. That is exactly double Opus 4.8 and the most expensive model Anthropic has priced for general use.

Do I need usage credits? Only if you want to keep using Fable 5 on a subscription after June 22, or if you exceed your plan's included usage. Credits are prepaid, enabled under Settings > Usage, and billed at standard API rates. Without them, your access reverts to your plan's included allowance.

How is Fable 5 priced versus Opus 4.8? Fable 5 is exactly twice the price: $10/$50 per million tokens versus $5/$25 for Opus 4.8. Inside subscription plans, Fable 5 also weighs roughly double the usage of Opus toward your limits. For most work, Opus 4.8 is the better cost-to-quality default; Fable 5 is the model to escalate to on long-horizon, complex tasks where it leads.