By AnthropicLearning Objectives
- Understand what Claude Fable 5 is and how it relates to Mythos 5 and Opus 4.8
- Evaluate Fable 5's capability and benchmark claims across coding, knowledge work, and science
- Explain the safeguard-and-fallback design that makes a Mythos-class model safe to release publicly
- Pick the right plan and pricing path for accessing Fable 5
What Is Claude Fable 5?
Claude Fable 5 is Anthropic's most capable publicly available model, released on June 9, 2026. It is the public, safeguarded version of Anthropic's Mythos-class frontier model — the same model line that, in its restricted form, powers Project Glasswing's cybersecurity work. Fable 5 supersedes Claude Opus 4.8 as the top generally available flagship.
Anthropic describes Fable 5 as "state-of-the-art on nearly all tested benchmarks of AI capability" across software engineering, knowledge work, vision, and scientific research. The model can work autonomously for extended periods, retaining memory across millions of tokens — Stripe reported that Fable 5 "compressed months of engineering into days" on a codebase migration.
💡Key Concept
Fable 5 vs. Mythos 5: They are the same underlying model. Fable 5 ships with safeguards on and is open to the public. Mythos 5 is the same model with safeguards lifted, restricted to vetted partners (Project Glasswing cybersecurity teams, with biomedical researchers next). See the Claude Mythos 5 page for the restricted-access story.
Capabilities and Benchmarks
Fable 5 posts the highest scores Anthropic has recorded across several independent evaluations:
| Benchmark | Result | Notes |
|---|---|---|
| FrontierCode (Cognition) | Highest-scoring model | Long-horizon software engineering |
| Core analytics benchmark | 90% | A 10-point improvement over Opus 4.8 |
| CursorBench | Top score | Agentic coding inside the editor |
| Hebbia Finance Benchmark | Top score | Knowledge-work over financial documents |
Beyond coding, Fable 5's vision is strong enough to reconstruct working code from a screenshot and to play games without auxiliary tools. In molecular biology, Anthropic says the model outperformed specialized protein models at predicting virus shell assembly, and that domain experts preferred its novel hypotheses roughly 80 percent of the time over Opus-class models.
Safeguards and the Fallback Design
Releasing a Mythos-class model publicly required a new safety architecture. Fable 5 runs three safety classifiers that, when triggered, fall back to Claude Opus 4.8 rather than answering directly. The classifiers cover:
- Cybersecurity and offensive hacking tasks
- Sensitive biology and chemistry research
- Model distillation attempts (using Fable 5 to train competing models)
Anthropic reports that roughly 95 percent of Fable 5 sessions involve no fallback at all — the safeguards are designed to catch a narrow band of high-risk requests rather than throttle everyday use.
⚠️Warning
The fallback can over-trigger. Because the classifiers are broad, security researchers have reported that routine requests — reading a security blog post, asking for a code review, or "write secure code" — sometimes trip the cybersecurity filter and downgrade the response to Opus 4.8. Anthropic points approved professionals to its Cyber Verification Program for fewer restrictions. This is the same tension explored on the Mythos 5 page.
Availability
📝Note
US export controls lifted (June 30, 2026): A mid-June US government directive — triggered after Amazon researchers found a jailbreak that could coax Fable 5 into producing cyberattack guidance — had forced Anthropic to pull Fable 5 and Mythos 5 offline for all foreign nationals, whether inside or outside the United States. On June 30, 2026 the Commerce Department lifted the controls. Commerce Secretary Howard Lutnick said the department spent two weeks working with Anthropic to analyze and approve the model, and Anthropic — which agreed to proactively detect and report security risks — began restoring public access on July 1, 2026. The separate identity-verification program (a government photo ID plus a live selfie, required for all Claude users from July 8, 2026) remains in place as a standing safeguard.
- Claude API and consumption-based Enterprise plans: available to US persons; access for foreign nationals was restored from July 1, 2026 following the June 30 lift of export controls
- Subscription plans (Pro, Max, Team, seat-based Enterprise): included at no extra cost for up to 50 percent of weekly usage limits through July 7, 2026; from July 8, 2026 access moves to usage credits at API rates. Anthropic says the change is driven by capacity constraints and is temporary — it aims to restore Fable 5 to standard subscription plans once capacity allows
- Mythos 5 (safeguards lifted): restricted to Project Glasswing cybersecurity partners, expanding to biomedical researchers, with a broader trusted-access program planned
Pricing
- Full Fable 5 access
- Millions-of-tokens memory
- Vision + agentic tools
- Free Fable 5 access through July 7, 2026, then usage credits
- Temporary, capacity-driven change
- Safeguards lifted for vetted partners
- Project Glasswing + biomedical access
Fable 5's per-token pricing ($10 input / $50 output per million tokens) sits above standard Opus 4.8 pricing ($5 input / $25 output) — the premium buys the frontier capability tier. For routine professional work where Opus 4.8 is sufficient, Opus remains the more economical default.
How Fable 5 Compares
| Model | Public Access | Positioning |
|---|---|---|
| Claude Fable 5 | Yes (API + subscriptions) | New public flagship; Mythos-class capability with safeguards |
| Claude Mythos 5 | No (vetted partners only) | Same model, safeguards lifted; cybersecurity + biomedical work |
| Claude Opus 4.8 | Yes (API + claude.ai) | Prior flagship; Dynamic Workflows; economical default for routine work |
| GPT-5.5 (OpenAI) | Yes | Native computer use; largest third-party ecosystem |
| Gemini 3.1 Pro (Google) | Yes | Strong multimodal; deep Workspace integration |
Strengths
- Frontier capability, publicly available — Mythos-class performance without the invite-only gate
- State-of-the-art benchmarks — leads FrontierCode, CursorBench, and the Hebbia Finance Benchmark; 90 percent on a core analytics benchmark
- Long-horizon autonomy — works across millions of tokens of memory; "months of engineering into days" on real migrations
- Strong vision and science — reconstructs code from screenshots; outperformed specialized protein models on virus-shell prediction
- Safe-by-default release — three safety classifiers fall back to Opus 4.8 on high-risk requests, with no fallback in roughly 95 percent of sessions
Limitations and Considerations
- Over-broad safeguard triggers — the cybersecurity classifier can downgrade legitimate security and code-review work to Opus 4.8; approved researchers need the Cyber Verification Program
- Premium pricing — $10 input / $50 output per million tokens is double standard Opus 4.8; not the right default for routine work
- Subscription access is temporarily usage-credit based — free on Pro, Max, and Team for up to 50 percent of weekly limits through July 7, 2026, then usage-credit based; Anthropic says this is a temporary, capacity-driven change and plans to restore standard subscription access when capacity allows
- Export-control saga resolved — a mid-June 2026 US directive briefly suspended Fable 5 and Mythos 5 access for non-US persons after a jailbreak was found; Commerce lifted the controls on June 30 and Anthropic restored access from July 1
- Closed model — API-only; no downloadable weights
- Mythos 5 stays gated — the unrestricted capabilities remain limited to vetted partners
Related Tools
- Claude Mythos 5 — The same model with safeguards lifted, restricted to Project Glasswing partners
- Claude Opus 4.8 — Anthropic's prior flagship and Fable 5's safeguard-fallback model
- Claude.ai — Anthropic's flagship chat interface, where subscribers reach Fable 5
- Claude Code — Anthropic's terminal coding agent, a primary surface for frontier-model work
Key Takeaways
- Claude Fable 5 is Anthropic's new public flagship (released June 9, 2026) — a safeguarded version of its Mythos-class model that supersedes Opus 4.8 as the top generally available model
- Fable 5 is state-of-the-art on nearly every capability benchmark Anthropic tested, works autonomously across millions of tokens, and posted top scores on FrontierCode, CursorBench, and the Hebbia Finance Benchmark
- Three safety classifiers fall back to Opus 4.8 on cybersecurity, sensitive biology/chemistry, and model-distillation requests; roughly 95 percent of sessions see no fallback, though the cybersecurity filter can over-trigger on routine work
- Pricing is $10 input / $50 output per million tokens; subscription plans get free access through July 7, 2026 (up to 50 percent of weekly limits), then usage credits — a temporary, capacity-driven change Anthropic expects to reverse
- Mythos 5 is the same model with safeguards lifted, restricted to vetted Project Glasswing and biomedical partners
