The Direct Answer: Claude Fable (now Claude Fable 5.1) is Anthropic's "Mythos-class" frontier model, positioned above Claude Opus 5. It is built for complex, multi-stage reasoning and reading charts or tables buried inside dense documents. Plan access is strictly split: on Max plans and Premium Team/Enterprise seats, Fable is included for up to 50% of weekly usage limits. On Pro and Standard seats, it requires pay-as-you-go usage credits ($10/M input, $50/M output). For everyday office work, Opus 5 or Sonnet 5 remain the practical default. Fable is a specialised tool reserved for difficult, high-stakes tasks.
Who this is for: Claude users evaluating whether Fable 5.1 is worth usage credits or a plan upgrade, and professionals handling heavy multi-document analysis.
Skip this if: You don't use Claude. This is Anthropic-specific and does not change everyday workflows in ChatGPT, Gemini, or Copilot.
Historical background: The June 2026 suspension and release timeline ↓
On June 12, 2026, three days after the original Fable 5 launch, Anthropic temporarily suspended access under a US export-control directive following a reported safeguard bypass. Anthropic retrained the safety classifier, and the model was restored worldwide on July 1 without nationality verification.
On July 20, 2026, Anthropic established the permanent plan split: Max and Premium seats retain included access (capped at 50% of weekly limits), while Pro and Standard seats pay usage credits. On September 19, 2026, Anthropic released Fable 5.1 as the current model under identical pricing and plan terms, alongside Opus 5.
What Claude Fable Is Built To Do
Anthropic positioned the Claude Fable series (now Claude Fable 5.1) in a tier called "Mythos-class," sitting above Claude Opus 5. Two specific capabilities make working professionals pay attention:
1. It sustains long, multi-stage work
Earlier models were built for short exchanges: you ask, it answers. Fable was built to work autonomously across extended periods: planning a task in stages, checking its own output, and executing across hours inside agent tools. For rewriting an email, that adds little value. For reconciling 14 quarterly spreadsheets into a single tracker, it changes the entire workflow.
2. It reads documents the way you do
Fable can parse charts, tables, and diagrams embedded directly inside PDFs and image files, not just the surrounding text. For professionals in finance, audit, legal, and analytics who deal with annexure-heavy reports, the model can reason about the table on page 37 rather than merely noting its presence.
How Access Works: Which Plan Gets Fable?
On the API, Fable 5.1 costs $10 per million input tokens and $50 per million output tokens, roughly double Opus 5. Within Claude subscriptions, access splits by plan tier:
- Max and Premium seats: Fable access is included at no extra charge, capped at 50% of your weekly usage limit.
- Pro and Standard seats: Fable access is not included in the monthly subscription fee. It runs entirely on pay-as-you-go usage credits from the first query.
Because of this structure, Fable 5.1 is designed as a precision tool for complex problems, not a daily replacement for Sonnet or Opus.
What This Means for Your Work
Nothing changes for your daily default. Opus 5 and Sonnet 5 handle drafting, summaries, status reports, and meeting prep without extra costs. Fable 5.1 is worth reaching for specifically when a task demands its twin strengths: autonomous multi-stage work, or parsing tables and charts buried inside dense PDFs.
If your week involves annexure-heavy PDFs, test Fable 5.1 against your real documents once. Then decide deliberately whether Max/Premium included quota or Pro usage credits are worth the cost. Try a disciplined prompt like this one, not a generic "summarize this":
I am uploading [N] documents/reports (sanitized — no client names or confidential figures). Work through them as one connected task: 1. Extract the key figures and claims from each document, INCLUDING anything that appears only in tables, charts, or diagrams 2. Build a single comparison table across all documents 3. Flag every inconsistency — places where two documents disagree on a number, a date, or a stated decision 4. End with: the 3 findings I would present to my reporting manager, in one line each If a chart or table is unclear, say which page and what is ambiguous — do not guess values.
For genuinely heavy PDFs, pair it with the step-by-step method in The PDF Cruncher: that workflow works on either model.
The Honest Take
The lesson from frontier model rollouts is clear: the model you can reliably open beats the model that's better on paper. Keep Opus 5 or Sonnet 5 as your stable default. Treat Fable 5.1 as a tool you reach for on difficult problems, not something you rebuild your whole operational stack around. The Copilot vs Claude vs Gemini comparison for everyday office tools stands unchanged.
Try it on one real document stack. Then decide deliberately whether Max/Premium included quota or Pro usage credits justify the expense for your workflow.
The Real World Story
I tested Fable on the most boring, painful task I had: cross-checking figures across a stack of quarterly review decks.
The decks had the usual problem: the same metric reported slightly differently in three places, with the real number buried in an appendix table. Previous models would summarize the slide text and miss the appendix entirely. Fable caught a mismatch between a chart on one deck and a table in another, and told me which page each came from. It was genuinely impressive.
That test proved where frontier models earn their keep. But it also proved why you need a stable fallback: when a top-tier model faces quota limits, credit costs, or policy shifts, having an everyday workflow anchored on Opus or Sonnet keeps your day job moving without interruption.