# Claude Fable 5 > [!abstract] What this note is > A profile of **Claude Fable 5**, the Anthropic model that powers this vault's Claude Code sessions — including the one that researched and wrote this note about itself, from sources rather than self-knowledge (its training ends before its own release). Fact-checked 2026-07-10 via a deep-research run: 99 agents, every claim verified by three independent adversarial checks, zero claims refuted; raw report in `.claude/research/`. Fast-moving subject — accurate as of the research date. ## What it is Claude Fable 5 (API model ID `claude-fable-5`) is the first model in Anthropic's **Claude 5 family**, announced June 9, 2026. Official docs call it "Anthropic's most capable widely released model," built for demanding reasoning and long-horizon agentic work. It introduces a new **Mythos-class** tier that sits above the Opus line — the announcement describes Fable 5 as "a Mythos-class model made safe for general use," and the migration docs treat Claude Opus 4.8 as its predecessor. ## Fable and Mythos The family launched as a pair: - **Claude Fable 5** — generally available, and includes safety classifiers that can decline requests in dual-use domains (cybersecurity, biology). - **Claude Mythos 5** (`claude-mythos-5`) — identical capabilities, specs, pricing, and API behavior, but *without* those classifiers. Not generally available: access is invitation-only through **Project Glasswing**, Anthropic's program offering the unclassified model to approved organizations (initially ~50 cyberdefenders and critical-infrastructure providers, in collaboration with the US government) for defensive cybersecurity work. The docs say Mythos "shares Claude Fable 5's capabilities without the safety classifiers." The stronger phrasing that they are *the same underlying model weights* is implied but only stated explicitly in third-party reporting *(single-source framing; unverified against official docs)*. ## Specs and pricing Per the [platform pricing docs](https://platform.claude.com/docs/en/about-claude/pricing), both models: **$10 / $50 per million input/output tokens** (exactly double Opus 4.8), a **1M-token context window** at standard pricing with no long-context surcharge, **128k max output tokens**, adaptive thinking always on, and a **January 2026 knowledge cutoff**. Available on the Claude API, AWS (Claude Platform and Bedrock), Google Cloud Vertex AI, and Microsoft Foundry. ## In Claude Code Fable 5 is the most capable model available in Claude Code but not the default on any account type — it's selected explicitly (`/model fable`, or the `best` alias, which resolves to Fable where the organization has access). When its cybersecurity/biology classifiers flag a request, Claude Code automatically re-runs it on the default Opus model and notes this in the transcript. ([docs: model configuration](https://code.claude.com/docs/en/model-config)) ## A bumpy first month Access to both models was suspended around June 12, 2026 — three days after launch — and restored around June 30–July 1 per Anthropic's "Redeploying Claude Fable 5" statement. The attribution of the pause to a US export-control directive appears only in third-party reporting *(unconfirmed by official sources)*. Docs assert full general availability as of this note's research date. > [!warning] Verification caveat > The announcement page itself (`anthropic.com/news/claude-fable-5-mythos-5`) refuses programmatic fetches, so its distinctive phrasings were confirmed through search snippets, the official AWS partner blog, and the near-verbatim mirror at [platform.claude.com](https://platform.claude.com/docs/en/about-claude/models/introducing-claude-fable-5-and-claude-mythos-5) — every load-bearing fact above was independently confirmed on official docs pages. ## Sources 1. [Introducing Claude Fable 5 and Claude Mythos 5 — platform docs](https://platform.claude.com/docs/en/about-claude/models/introducing-claude-fable-5-and-claude-mythos-5) 2. [Models overview — platform docs](https://platform.claude.com/docs/en/about-claude/models/overview) 3. [Pricing — platform docs](https://platform.claude.com/docs/en/about-claude/pricing) 4. [Migration guide — platform docs](https://platform.claude.com/docs/en/about-claude/models/migration-guide) 5. [Model configuration — Claude Code docs](https://code.claude.com/docs/en/model-config) 6. [Announcement — anthropic.com](https://www.anthropic.com/news/claude-fable-5-mythos-5) *(confirmed via mirrors; page blocks direct fetch)* 7. [Project Glasswing — anthropic.com](https://www.anthropic.com/glasswing) 8. [Redeploying Claude Fable 5 — anthropic.com](https://www.anthropic.com/news/redeploying-fable-5) ## Related - [[Claude Code Skills]] — the skills available to the sessions this model runs - [[Claude Code Integration]] — how this vault doubles as a Claude Code project