Model Comparison
Fable 5 vs Claude Opus: What's Actually Different
Two flagships, one company, built for genuinely different things.
Fable 5 and Claude Opus 4.8 are both Anthropic flagships, but they represent different bets about what top-tier AI should do. Fable 5 is the first model in the new Mythos class, designed for long-horizon autonomous work where performance scales with task length. Opus 4.8 is the proven generalist at the top of the Claude 4 family. This page covers what separates them, where they cooperate, and how to choose.
Fable 5 (released June 9, 2026) is Anthropic's first Mythos-class model, optimized for long autonomous tasks where performance scales with task length; it costs $10 per million input tokens and $50 per million output tokens. Claude Opus 4.8 is Anthropic's flagship generalist in the Claude 4 family. The two models cooperate: Fable 5 automatically routes classifier-flagged sensitive queries in cybersecurity, bio/chem, and distillation to Opus 4.8, and more than 95% of Fable 5 sessions involve no such fallback.
Side by Side
How Fable 5 and Claude Opus 4.8 differ
Both models come from Anthropic, but they are optimized for different kinds of work.
Fable 5: Mythos class
The first model in Anthropic's Mythos class, released June 9, 2026. Designed for multi-step autonomous tasks where performance increases as the task grows longer.
Claude Opus 4.8: generalist flagship
The top model in the Claude 4 family. A broad-capability generalist built for demanding reasoning, writing, analysis, and coding tasks across session lengths.
Task length sweet spot
Fable 5 is purpose-built for long-horizon runs where its performance scales with task length. Opus 4.8 excels across both short and extended sessions without that specific design bias.
Fable 5 API pricing
Fable 5 is priced at $10 per million input tokens and $50 per million output tokens. Consult Anthropic's pricing page for current Opus 4.8 rates.
Built-in safety routing
Fable 5 includes a classifier that detects sensitive queries and automatically routes them to Claude Opus 4.8. More than 95% of sessions involve no such handoff.
Different model lines
Fable 5 introduces the Mythos class, a new line separate from the Claude family. Opus 4.8 sits within the established Claude 4 family alongside Sonnet and Haiku.
How They Cooperate
Safety routing: when Fable 5 hands off to Opus
One of the less obvious aspects of Fable 5 is that it does not always operate alone. A built-in classifier monitors queries for signals in three sensitive domains: cybersecurity, biological and chemical information, and knowledge distillation. When the classifier fires, Fable 5 automatically routes the query to Claude Opus 4.8 rather than answering directly.
Anthropic reports that more than 95% of Fable 5 sessions involve no fallback, meaning the handoff is designed for edge cases rather than everyday use. For most autonomous tasks, Fable 5 handles the full run without touching Opus 4.8.
This cooperation model matters for developers building on the API. If your application involves any of the flagged domains, you may encounter Opus 4.8 responses from a Fable 5 endpoint. Knowing the routing logic exists helps you interpret latency differences and response characteristics when they appear.
Which to Choose
Picking the right model for your task
Choose Fable 5 when you are running long autonomous pipelines: multi-step research, extended coding agents, or iterative tasks where the model works largely unsupervised over many turns. Fable 5 is designed on the assumption that performance should improve as the task grows, which makes it a natural fit for agentic workloads.
Choose Claude Opus 4.8 when you need a broad generalist for demanding one-shot or moderate-length sessions: complex analysis, nuanced writing, multi-domain reasoning, or tasks that cross into the sensitive domains Fable 5 would route away from. Opus 4.8 handles the full range without a classifier handoff.
For teams building products, the two models are not mutually exclusive. Fable 5 can orchestrate a long task while automatically calling Opus 4.8 for sensitive subtasks. Understanding both models' strengths lets you design workflows that use each where it performs best.
Learn to work with Claude's flagship models, hands-on
Reading a comparison and actually knowing how to prompt, chain, and deploy these models are different skills. The free Claude Academy curriculum is practical from the first lesson, no theory-only modules.
Frequently Asked Questions
When did Fable 5 come out?
Fable 5 was released on June 9, 2026. It is the first public model in Anthropic's Mythos class, a new model line separate from the Claude 4 family.
What does Mythos class mean?
Mythos is a new model class Anthropic introduced with Fable 5. Unlike the Claude family, which is organized by capability tier (Opus, Sonnet, Haiku), the Mythos class signals a different design philosophy: performance that scales with task length, optimized for long-horizon autonomous work rather than general-purpose use.
What is Fable 5's API pricing?
Fable 5 is priced at $10 per million input tokens and $50 per million output tokens through Anthropic's API. These figures come from Anthropic's official release announcement at anthropic.com/news/claude-fable-5-mythos-5.
Why does Fable 5 route some queries to Claude Opus 4.8?
Fable 5 includes a built-in classifier that detects queries in sensitive domains: cybersecurity, biological and chemical information, and knowledge distillation. When the classifier flags a query, Fable 5 routes it to Opus 4.8 rather than answering directly. This is a safety design. More than 95% of sessions involve no such routing.
Is Fable 5 better than Claude Opus 4.8?
They are optimized for different things, so the answer depends on the task. Fable 5 is built for long autonomous tasks where performance scales with task length. Claude Opus 4.8 is a broad generalist that handles demanding tasks across session lengths and covers sensitive domains directly without routing. For sustained agentic pipelines, Fable 5. For complex one-shot or moderate-length work, Opus 4.8.
Can I use Fable 5 and Claude Opus 4.8 together?
Yes, and Fable 5 already does this automatically via its safety routing classifier. You can also design your own pipelines that use Fable 5 for long-horizon orchestration and explicitly call Opus 4.8 for subtasks that fall into sensitive domains or that benefit from a generalist approach.
What sensitive domains trigger the Fable 5 handoff?
Fable 5's classifier detects three categories: cybersecurity, biological and chemical information, and knowledge distillation. Queries matching these are automatically routed to Claude Opus 4.8. The handoff is transparent and requires no action from the developer for it to work.
Is Fable 5 available on the Anthropic API now?
Fable 5 was publicly released on June 9, 2026. Check Anthropic's official documentation and console for current API availability, model IDs, and any plan or region restrictions.
Put both models to work
The best way to understand the difference is to use them.