AI Tools

Claude Fable 5.1 vs Claude Mythos 5.1: what's the difference?

Published 2026-09-02 · Researched and price-checked on official sources

Claude Fable 5.1 vs Claude Mythos 5.1: Anthropic announced both in September 2026 as the same underlying model with different levels of safeguards. Fable 5.1 is generally available at $10 per million input and $50 per million output tokens. Mythos 5.1 stays restricted to vetted cybersecurity and life-science organizations.

Key facts

What matters

Claude Fable 5.1 and Claude Mythos 5.1 are the same model; Mythos ships with fewer safeguards for vetted cyberdefense and life-science teams, not a separate, stronger model, per Anthropic's announcement.
Fable 5.1 API pricing is $10 per million input and $50 per million output tokens, with cache reads cut 75% to $0.25 per million tokens, lowering typical workload cost about 25% versus Fable 5 and up to around 45% on highly agentic work.
Fable 5.1 is live today on the Claude API as claude-fable-5-1 plus AWS, Google Cloud, and Azure; Mythos 5.1 is currently limited to US organizations in Anthropic's trusted access programs.
On Max plans Fable 5.1 is included up to 50% of your weekly usage limits; on Pro plans it runs on pay-as-you-go usage credits billed at standard API rates, per Anthropic's support article.
Fable 5.1 beats Fable 5 and Opus 5 on Anthropic's headline benchmarks (52.6% vs 24.7% and 29.0% on Terminal-Bench-Science 0.1) and also edges GPT-5.6 Sol where both were measured.

What are Claude Fable 5.1 and Claude Mythos 5.1?

Claude Fable 5.1 is Anthropic's flagship model for coding, knowledge work, and long-running agent tasks, announced in September 2026 as the successor to Fable 5. Anthropic's pricing page calls the family "next generation intelligence for long-running agents," and the announcement describes the pair as "the world's most advanced models for coding and knowledge work."

Claude Mythos 5.1 is not a separate model. Anthropic states plainly that Fable 5.1 and Mythos 5.1 are the same model with different levels of safeguards. Fable 5.1 is generally available; Mythos 5.1 is only available through trusted access programs, with safeguards designed for cybersecurity and life-science work.

Fable 5.1 is available today on all platforms including Amazon Web Services, Google Cloud, and Microsoft Azure, and developers can call it as claude-fable-5-1 on the Claude API. Anthropic notes Fable 5.1 defaults to High effort in Claude Code and to Medium effort in Claude Cowork and on Claude.ai.

Claude Fable 5.1 vs Claude Mythos 5.1: what actually differs?

The difference is guardrails, not weights. Anthropic's own cost-curve charts for Terminal-Bench 4.0 show Mythos 5.1 ahead of Fable 5.1, and the company explains the gap "reflects the tasks on which our earlier, less precise cyber safeguards intervened." With the safeguard improvements released today, Anthropic expects the difference between the two models to become much smaller.

For security teams, the practical change is fewer false refusals. Anthropic says its newest safeguards block 60% fewer false positives in cybersecurity, and that Fable 5.1 can now be used to discover software vulnerabilities, though not to develop exploits for them.

Enterprises get a privacy option in the same announcement: the Enterprise Frontier Safeguards (EFS) system stores data in cloud infrastructure the customer controls, giving complete privacy equivalent to zero data retention. EFS rolls out to enterprise customers in phases starting later this fall. Until it is available, eligible customers can run Fable 5.1 with zero data retention.

What does Claude Fable 5.1 cost on the API?

Fable 5.1 keeps Fable 5's headline prices: $10 per million input tokens and $50 per million output tokens. The cost story is prompt caching. Cache reads, where the model reuses context it has already processed, drop 75% to $0.25 per million tokens on Fable 5.1, down from $1.00 on Fable 5. Cache writes stay at $12.50 per million tokens.

Anthropic estimates that cuts the cost of typical workloads by around 25% versus Fable 5, and by up to around 45% for complex coding and highly agentic work, where cache reads dominate the bill. The company also says that at Low or Medium effort, Fable 5.1 matches or beats Fable 5's results at much lower cost.

Fable 5.1 is more expensive per token than Opus 5 ($5 input, $25 output), so the cheaper agentic economics only materialize when prompt caching applies. Teams that keep context small and calls short may still pay more on Fable 5.1 than on Opus 5, which is worth pricing out before migrating traffic.

How does Fable 5.1 benchmark against Fable 5, Opus 5, and GPT-5.6 Sol?

Anthropic's September 2026 benchmark table shows Fable 5.1 leading its own family and GPT-5.6 Sol on the categories where all were measured:

Terminal-Bench-Science 0.1 (agentic scientific research): Fable 5.1 scores 52.6%, versus 24.7% for Fable 5, 29.0% for Opus 5, and 22.4% for GPT-5.6 Sol.
Humanity's Last Exam with tools (multidisciplinary reasoning): 65.0% for Fable 5.1, 63.8% for Fable 5, 63.6% for Opus 5. No GPT-5.6 Sol figure is shown.
CursorBench 3.2.0 (agentic coding): 73.4% for Fable 5.1, versus 70.5% (Fable 5), 70.0% (Opus 5), and 67.2% (GPT-5.6 Sol).
OSWorld 2.0 (computer use, partial credit): 77.9% for Fable 5.1, 72.9% for Fable 5, 75.4% for Opus 5; strict scores are 41.7%, 36.1%, and 39.6%. No GPT-5.6 Sol number is shown.
GDPval-AA v2 (knowledge work): Fable 5.1 scores 1853, ahead of Opus 5 at 1824, Fable 5 at 1723, and GPT-5.6 Sol at 1711.

Read those numbers with one caveat Anthropic prints in the same post: all scores were measured with production safeguards enabled, and on tasks where safeguards intervened, Fable 5.1 and Fable 5 scored zero on OSWorld 2.0 and Fable 5 scored zero on AutomationBench. The Terminal-Bench-Science 0.1 standard error is plus or minus 3.5 to 4.5 points per model.

Who can use Claude Mythos 5.1, and who cannot?

Mythos 5.1 is not on sale. Anthropic says it is available to vetted cyberdefenders and life scientists, and currently only to a set of US organizations, with the company coordinating with the US government to expand access to more domestic and international partners. Cyberdefense teams register interest through Anthropic's CVP access program; the biology program, built in partnership with the US government, expects to open enrollment for scientists soon.

Pro and Max subscribers, Claude Code users, and API customers all get Fable 5.1, not Mythos. If your use case is defensive security work on sanctioned systems, Mythos is the model whose safeguards are tuned for it, but it is gated behind organizational vetting rather than a payment plan.

One regulatory note applies to the whole release: as a model released after August 2, 2026, Fable 5.1 falls under the EU AI Act Code of Practice on transparency of AI-generated content that Anthropic signed. Outputs carry an invisible watermark with no practical impact on quality, and Anthropic is rolling out a watermark detection API in private preview for regulators, media, and obligated enterprises.

How does Fable 5.1 work on Claude Pro and Max plans?

Anthropic's support article confirms Fable 5 and Fable 5.1 are available on all paid plans: Pro, Max, Team, and Enterprise. The billing differs by tier, and that difference is the real cost catch for subscribers.

On Max plans, and on premium seats on Team and legacy seat-based Enterprise plans, Fable 5.1 is a standard part of the plan. You can use up to 50% of your weekly usage limits on Fable models at no extra cost, and Fable usage draws down the weekly limit faster than other Claude models.

On Pro plans, and on standard Team and Enterprise seats, Fable 5.1 is not included in your plan usage limits. It runs on pay-as-you-go usage credits from the start, billed at standard API rates, which means the $10 and $50 per million token prices apply on top of the $20 monthly Pro subscription. A one-time promotional credit was given when Fable 5 moved to usage credits in July 2026; Anthropic confirms there is no equivalent credit for Fable 5.1.

Pro costs $20 per month, or $17 per month billed annually at $200 up front. Max starts at $100 per month and lets you choose 5x or 20x the usage of Pro. For teams that live in Claude Code, where Fable 5.1 defaults to High effort, the Max route is usually the cheaper one once Fable-heavy weeks start; light subscribers should price the usage credits before upgrading. Claude Code version 2.1.250 or later is required to select Fable 5.1.

At a glance

Model	Input	Output	Cache read	Cache write
Claude Fable 5.1	$10	$50	$0.25	$12.50
Claude Fable 5	$10	$50	$1.00	$12.50
Claude Opus 5	$5	$25	$0.50	$6.25
Claude Sonnet 5	$2	$10	$0.20	$2.50
Claude Haiku 4.5	$1	$5	$0.10	$1.25

Claude API pricing per million tokens, Anthropic official pricing page, verified September 2, 2026

FAQ

Is Claude Fable 5.1 free with Claude Pro?

No. Anthropic's support article says Fable 5.1 is available on all paid plans, but on Pro plans it runs on pay-as-you-go usage credits billed at standard API rates from the start. Only on Max plans, and on premium Team and Enterprise seats, is Fable 5.1 included as a standard part of the plan, up to 50% of your weekly usage limits.

Can I use Claude Mythos 5.1?

Only if your organization is a vetted cyberdefender or life-science team admitted to Anthropic's trusted access programs, and access is currently limited to US organizations. Cyber teams register interest through Anthropic's CVP program, and the biology access program, developed with the US government, plans to open enrollment for scientists soon. General subscribers and API customers get Fable 5.1 instead.

Is Claude Fable 5.1 cheaper than Claude Fable 5?

Yes for token-billed usage. Input and output prices are unchanged at $10 and $50 per million tokens, but cache reads dropped 75% to $0.25 per million tokens, so Anthropic estimates typical workloads cost around 25% less and highly agentic workloads up to around 45% less than Fable 5.