Fable 5 usage limits, tracked from day one
Anthropic's new flagship tier ships with its own weekly usage window, reported through a brand-new API format. tokenkarma supported it the day it dropped, so your most capable model is also your most visible one.
Fable 5, the first Claude 5 family model, is metered on its own per-model weekly window on Claude plans, separate from your usual session and weekly caps. And as of July 2, 2026 it is also capped: through July 7, paid plans get Fable 5 for up to 50% of their weekly usage limits at no extra cost.
The launch also changed Anthropic's usage API: new models report their windows in a new format, which older trackers simply cannot see. tokenkarma shipped support the same day, and shows your Fable 5 window as a plain percentage with a reset countdown.
Right now the number that matters is 50: as of July 2, 2026, Fable 5 is back on paid Claude plans, capped at 50% of your weekly usage limits through July 7.
Fable 5 is back: the 50% weekly cap, explained
On July 1, 2026, Anthropic brought Fable 5 back to Claude subscriptions after pulling it from plans in late June. Through July 7, users on Pro, Max, Team and premium Enterprise plans can spend up to 50% of their weekly usage limits on Fable 5 at no extra cost. Past that ceiling, or after the window closes, using Fable 5 takes usage credits, or you switch back to another Claude model. Anthropic has said it may extend the window if capacity allows.
| Date | Fable 5 on Claude plans |
|---|---|
| June 9, 2026 | Ships, included in paid plans |
| June 23, 2026 | Removed from plans, usage credits required |
| July 1 to 7, 2026 | Back, up to 50% of weekly limits at no extra cost |
| After July 7 | Usage credits again, unless Anthropic extends |
The catch: by Anthropic's own guidance, Fable 5 burns your limits roughly 2x faster than Opus. Half of your weekly window goes quicker than it sounds, and the only way to pace yourself is to watch the number while you work. That is exactly what the Fable 5 row in tokenkarma is for.
What Fable 5 is, and why it gets its own limit
Fable 5 is Anthropic's most capable generally available model: the first of the Claude 5 family, in a new Mythos-class tier that sits above Claude Opus. Its sibling Mythos 5 is the same underlying model, offered to approved organizations. Flagship capability means flagship compute, so Anthropic meters Fable 5 separately instead of folding it into your general usage.
In practice that means a Claude plan now carries one more number to watch: your 5-hour session, your overall weekly window, and a Fable 5 weekly window of its own. Run your heavy reasoning on Fable 5 all Monday and your regular Claude usage can look fine while the flagship window quietly empties.
The Fable 5 row in tokenkarma
This is the exact fuel bar the tokenkarma island shows for Fable 5: the live percentage of your weekly window and the reset countdown, with an identity you can spot at a glance.
The real component from the Mac app, not a mockup: same grid, same 5px bar, its own violet identity. Every other source keeps the standard green-to-red scale.
How the Fable 5 weekly window works
Anthropic's usage API reports Fable 5 as a per-model window on a 7-day cycle. You draw it down as you use the model, and it comes back on a weekly rhythm rather than every few hours, so a heavy day costs you headroom for the rest of the week.
| Window | Scope | Refills |
|---|---|---|
| 5-hour session | All Claude usage | Every 5 hours |
| Weekly window | All Claude usage | On a 7-day cycle |
| Fable 5 window | Fable 5 only | On its own 7-day cycle |
Anthropic does not publish the exact quota behind the percentage, and it differs by plan. That is why a live reading matters more than a rule of thumb: the API tells you where you stand right now, and tokenkarma keeps that number in your menu bar.
New model, new API format, day-one support
The Claude 5 launch changed how Anthropic reports usage: new models publish their windows in a new API format that older integrations do not read. A tracker built for the old format keeps working for Opus or Sonnet and silently shows nothing for Fable 5, exactly when you start leaning on the most expensive tier.
tokenkarma shipped support for the new format the day Fable 5 dropped. That is the standard we hold the tracker to: when a new model lands in your stack, it lands in your island the same day, with its window, its reset time and its alerts.
How to see your Fable 5 usage
Claude's own interface tells you when a window is exhausted, but it does not put the flagship's weekly number in front of you while you work. tokenkarma does:
- a dedicated Fable 5 row in the Mac menu bar island and the web dashboard,
- the live percentage of the weekly window, with the reset countdown,
- alerts at the thresholds you choose, before the wall and not after,
- your Fable 5 usage next to Claude, ChatGPT, Gemini, Grok and Cursor, one glance for the whole stack.
Track Fable 5 next to every other AI you pay for
A flagship window is one number in a bigger picture: most Fable 5 users also juggle Claude's regular windows, a ChatGPT plan and an IDE like Cursor. tokenkarma puts every limit in one view so you can spend your most capable model where it counts. See the full AI usage tracker, the side-by-side limit comparison, or how Claude's 5-hour and weekly windows work.
Frequently asked questions
What is Fable 5?
Fable 5 is Anthropic’s new flagship Claude tier, the first model in the Claude 5 family and part of a new Mythos-class tier that sits above Claude Opus in capability. It is generally available on Claude plans; Mythos 5 is the same underlying model offered separately to approved organizations. Because it is the most capable tier, it gets its own usage accounting.
Does Fable 5 have its own usage limit?
Yes. On Claude plans, Fable 5 reports its own per-model weekly window, separate from your overall session and weekly windows. Anthropic does not publish the exact quota as a hard number, it varies by plan and can change. tokenkarma reads the live state and shows it as a plain percentage with the reset time.
When does the Fable 5 limit reset?
Fable 5 usage is metered on a 7-day window per model, so your headroom comes back on a weekly cycle rather than every few hours. The exact reset moment depends on when your window opened. tokenkarma shows the countdown next to the bar, for example "reset in 3d".
Why is Fable 5 limited to 50% of my weekly usage?
Anthropic brought Fable 5 back to Claude subscriptions on July 1, 2026 with a ceiling: through July 7, Pro, Max, Team and premium Enterprise users can spend up to 50% of their weekly usage limits on Fable 5 at no extra cost. Past that ceiling, or after July 7, it takes usage credits, or you switch to another Claude model. Anthropic has said it may extend the window if capacity allows. Keep in mind Fable 5 burns limits roughly 2x faster than Opus, so the 50% empties quicker than it sounds. This reflects Anthropic’s announcements as of July 2, 2026 and can change.
Why did my usage tracker miss Fable 5?
When the Claude 5 family launched, Anthropic changed its usage API: new models report their windows in a new format. Tools that only read the old format cannot see Fable 5 usage at all. tokenkarma shipped support for the new format the same day the model dropped, so the Fable 5 window shows up like any other source.
Can I track my Fable 5 usage in real time?
Yes. tokenkarma shows Fable 5 as its own row in the Mac menu bar island and on the web dashboard: a live percentage of the weekly window, the reset countdown, and alerts before you hit the cap. The row has its own violet identity so you can spot your flagship usage at a glance next to Claude, ChatGPT, Gemini, Grok and Cursor.
Is tokenkarma affiliated with Anthropic?
No. tokenkarma is independent and not affiliated with Anthropic. The limit mechanics described here reflect what Anthropic’s own usage API reports as of July 2026 and can change.
Know your Fable 5 window before it knows you
tokenkarma shows your Fable 5 weekly usage live in your menu bar, with a reset countdown and alerts before you hit the cap.
Works in your browser and on your Mac.
Not affiliated with Anthropic. Limit mechanics reflect what Anthropic's usage API reports as of July 2026 and can change.