Fable 5.2 launch offers: check your plan before running agents

Claude's Fable 5.2 launch includes temporary usage offers, but they are not unlimited agent access. Check your plan, reset expiry and credit settings before starting a long run. The short Pro/standard-Team window ends 17 November 2026 at 15:59 Singapore time.

Checked 12 November 2026. This follows up the real 9 November launch, not a claim that Anthropic announced a new offer today. It is separate from our Claude Max value comparison: the decision here is how to use a dated offer without mistaking it for permanent funding.

Which offer should you check for your plan?

Anthropic's launch-offer rules distinguish these groups:

Plan or seat	Offer and Singapore expiry
Pro, standard Team	Fable included up to 50% of weekly limits; eligible reset and inclusion end 17 November 2026, 15:59
Max, Ultra, premium Team, premium legacy seat-based Enterprise	One to three resets by paid tenure; expiry 9 March 2027, 15:59

Reset billing cutoff: before 10 November 2026, 02:00 Singapore; organization and seat conditions apply. Trials and excluded Enterprise categories need the full eligibility check. After Pro/standard-Team inclusion ends or its allowance is spent, enabled usage credits can fund Fable. All times above are converted from Anthropic's Pacific deadlines.

Before planning work, open the linked rules and match the exact seat, not just the organization's name. Save the offer and expiry shown in your own account. If those differ from a remembered announcement, resolve the account-specific question before relying on the allowance. This is a planning checklist, not evidence that every reader received a reset.

What does a Fable reset restore?

The offer restores weekly Fable consumption, not other weekly model usage; it also refills a five-hour window where present. It does not increase limits. The same offer page contains the full rules.

Anthropic's reset instructions direct you to Settings, Usage on web, Desktop or mobile. The button is not currently in Claude Code's terminal or IDE. Resets apply across the account, do not refund billed extra usage and do not change the credit balance.

For your own record, keep the reset type, expiry, usage before application and usage afterward. Do not combine those observations into a claim about tokens per reset. A percentage view is useful for tracking your account, not a universal conversion table for someone else's workload.

Does the offer make Fable available in every connected app?

No. Model entitlement and app integration are different checks. Anthropic's Claude Code legal page permits end users to sign into the unmodified Claude Code binary under its conditions, including hosted setups. It does not permit third-party collection or intermediation of Claude.ai credentials.

Use the Claude app directory to begin researching a supported workflow. A listing is not proof that its current model picker exposes Fable 5.2. Check the app's own release notes, installed runtime and selected account before starting. For a workspace such as T3 Code, verify the particular provider session rather than assuming every session in the same window shares one payer.

Do not paste a session token into a plugin to obtain the offer. If the official runtime cannot authenticate, diagnose that first with our agent login guide. A workaround that changes authentication also changes the question this offer page answers.

Should you switch all agent work to Fable 5.2?

Anthropic's 9 November announcement positions the models for demanding technical work and recommends starting most coding and knowledge tasks with Opus 5.5. That is provider guidance, not our own benchmark result.

Choose one bounded task with a clear acceptance test before moving a backlog. Keep the same input and review standard you use for the current model. Record whether the result passes, what manual correction it needs and the account usage shown. Do not infer a savings percentage from one successful demo or a social screenshot.

Avoid starting a long unattended job immediately before an offer deadline. Leave time to inspect its output and decide whether any continuation is worth separate credit use. A temporary allowance is most useful when you can tell what accepted work it produced.

What should you recheck before the offer expires?

Review model selection, current allowance, remaining resets and any enabled credit funding. Decide in advance whether to stop, choose another included model or authorize further spending. Recheck the live offer terms when publishing or reading this near expiry.

Keep model funding separate from the app's own fees and external tools. Our self-hosted-agent guide covers operating responsibilities; the subscription-versus-API guide covers the broader cost decision. Neither turns this launch week into a permanent promise. Treat the dates above as an offer window to verify, not a reason to remove spending controls.