Claude Fable 5.1

Claude Fable 5.1 is Anthropic's generally available Fable-class model, released September 1, 2026 alongside Claude Mythos 5.1. Anthropic says the two names are the same underlying model with different safeguards: Fable 5.1 is what most developers get; Mythos 5.1 is gated for vetted cybersecurity and life-science organizations. The product page at anthropic.com/claude/fable now leads with 5.1. API id is claude-fable-5-1.

Compare Claude Fable 5 if you are still on the June 2026 SKU, Claude Opus 5 if you want a cheaper broadly available frontier model, or GPT-6 Astra if you want OpenAI's competing flagship.

Model Specifications

US-only inference is available at 1.1x input and output pricing. Treat the 1M context figure as the Fable-class window Anthropic documents for this line; recheck the platform docs if you need a hard token cap.

Key Features

• Same list price, cheaper agents: input and output stay $10 / $50. Cache reads drop to $0.25. Anthropic estimates about 25% lower cost on typical workloads and up to about 45% on highly agentic jobs versus Fable 5, measured on August 2026 usage.
• Coding and long-running work: Anthropic reports Terminal-Bench 4.0 at 55.8% (60.9% on Mythos 5.1), CursorBench 3.2.0 at 73.4%, Terminal-Bench-Science 0.1 at 52.6%, and AutomationBench at 31.4%. Those are vendor evals with production safeguards on.
• Computer use: OSWorld 2.0 partial 77.9% and strict 41.7% on the authors' August 2026 task release. Anthropic warns these numbers are not comparable to older OSWorld 2.0 publications.
• Tighter safeguards, fewer false positives: cyber safeguards block about 60% fewer false positives than Fable 5. Fable 5.1 may identify software vulnerabilities but is still blocked from developing exploits. Biology R&D queries still route to Opus. You are not billed Fable prices for rerouted turns.
• Enterprise Frontier Safeguards: EFS stores monitoring data in customer-controlled cloud infrastructure. Until it rolls out, eligible customers can use Fable 5.1 with zero data retention. Default consumer path still requires 30-day retention for safety monitoring.

Limitation: Reddit and Claude Code threads after launch reported a 50% weekly Fable cap and a prompt-caching bug that burned limits (later noted as fixed in Claude Code 2.1.260). Recheck current caps in your plan. Do not treat launch-week complaints as frozen product facts.

Use Cases

• Multi-day coding in Claude Code: features that span a repo, review, and unattended verification loops.
• Knowledge work with files and vision: decks, PDFs, diagrams, and checking UI output against a design.
• Agent handoff: Cowork backlogs, Claude Tag in Slack (beta), browser use, and managed agents on the Claude Platform.

Getting Started

1. On Claude.ai or Claude Code, pick Fable 5.1 if your plan includes it (Pro, Max, Team, Enterprise).
2. On the API, set the model to claude-fable-5-1. Configure the Fallback API so cyber and bio routing works.
3. Prefer this SKU over Fable 5 for new agent sessions so you pick up the cache-read price cut.
4. Prototype here before requesting Mythos 5.1.

First-party resource: Introducing Claude Fable 5.1 and Claude Mythos 5.1.

Frequently Asked Questions

Is Fable 5.1 the same as Mythos 5.1?

Same weights, different safeguards. Fable 5.1 is generally available. Mythos 5.1 is invite-only for verified cyber and life-science work.

Should I still use Fable 5?

Only if a tool or contract pins the older id. For new work, 5.1 is the live Fable SKU on Anthropic's Fable page.

Why did my cyber question come back from Opus?

Safeguards routed it. You should not be charged Fable rates for that turn. Dual-use tasks such as exploit generation still go to Opus.

Alternatives

• Claude Mythos 5.1: same model, relaxed safeguards, trusted access only.
• Claude Fable 5: June 2026 generally available Fable SKU.
• GPT-6 Astra: OpenAI flagship at the same $10 / $50 list price.
• Claude Opus 5: lower list price, broader availability.

Tips

1. Budget for cache reads. That is where 5.1 actually gets cheaper than 5.
2. Keep Fallback API configured so a safeguard hit does not fail the request.
3. Measure cost per task, not list price. Anthropic argues 5.1 can undercut Opus on agent jobs because it uses fewer tokens; verify on your traces.

Conclusion

Claude Fable 5.1 is the public Fable-class model to use after September 1, 2026: same $10 / $50 sticker as Fable 5, much cheaper cache reads, and the sibling of Mythos 5.1. Start at the Fable product page or the launch post.