Claude Fable 5.2 Is Now Live on Emergent

Claude Fable 5.2, Anthropic's model for the most difficult technical work, is now available on Emergent. You can select it when you start a new project. Anthropic released Fable 5.2 on November 9, 2026, and reports that it produces output up to 30% faster than Fable 5.1 while posting large gains on agentic coding and long-running tasks.

If you already build with Fable 5.1 on Emergent, 5.2 is a substantial upgrade. It scores far higher on agentic coding and long-running work, runs faster, and blocks fewer harmless requests. It also sits alongside Claude Sonnet 5.5 and Opus 5.5 in your model picker as the top tier for builds that need the most reasoning.

How to use Fable 5.2 on Emergent

Choose Claude Fable 5.2 from the model selector in the prompt bar when you start a project. Model choice is set per project, so pick it before you describe your app.

Access runs through the Universal LLM Key, so there is no separate Anthropic account or API key to manage. Usage draws from your Emergent credits, the same way it does for every other supported model.

The rest of the flow is unchanged. Describe what you want in plain language, and the agent plans, builds, tests, and deploys the app from that conversation.

What Fable 5.2 brings to a build

Long, multi-step work is where Fable 5.2 stands apart. In Anthropic's testing, a team of agents running an earlier version of the model migrated a production service from Python to Rust, and the new service starts about 26 times faster. In another experiment, Fable 5.2 coordinated more than 1,000 subagents over eight days to build a bootable prototype operating system from an empty folder.

Your app won't need an operating system. But the same staying power applies when a build spans a database, user roles, payments, and several connected screens.

Speed improved too. Beyond the faster output, Anthropic says Fable 5.2 uses fewer tokens than Fable 5.1 at Medium effort, the default setting in Claude's own apps.

Early testers also point to more honest answers. Cognition reported that the model "says it doesn't know things instead of hallucinating." Oracle found that in a SQL test where two other models returned totals five to eight times too high, Fable 5.2 returned the expected results. For an app that calculates invoices or reports, that kind of accuracy matters more than raw speed.

Fable 5.2 makes big gains over Fable 5.1

Agentic coding and long-running work show the largest jumps from the previous version. The figures below come from Anthropic's launch post.

All scores are vendor-reported by Anthropic, with safeguards on, at each model's highest available effort unless noted. Fable 5.2 and 5.1 FrontierCode scores are at Xhigh effort; the Fable 5.2 Remote Labor Index score is at High effort. Scores as of November 2026.

Fable 5.2 beats its predecessor on every benchmark in the table. Remote Labor Index shows the biggest change: the score more than doubled, from 17.9% to 41.3%. AutomationBench rose 20 points, and FrontierCode 2.0 rose nearly 17 points to 49.1%, the highest score Anthropic recorded for any released model it tested.

The new model also leads Opus 5.5 on most of these benchmarks, though not all. Opus 5.5 rates slightly higher on AA-Briefcase Elo, and Google's Gemini 4 Carbon edges Fable 5.2 on AutomationBench, 53.9% to 51.4%.

Should you switch from Fable 5.1?

Yes, for any new build where you would have picked Fable 5.1. On top of the benchmark gains, Fable 5.2 produces output up to 30% faster. Anthropic also says it uses fewer tokens than Fable 5.1 at Medium effort, the default setting in Claude's own apps, which should lower the credits each task uses.

Fable 5.2 also refuses fewer harmless requests. Anthropic says it triggers about 55% fewer unnecessary safeguard blocks than Fable 5.1 while keeping the same protections.

Fable 5.2 vs. Mythos 5.2: which one is on Emergent

Fable 5.2 and Mythos 5.2 share the same underlying model. The difference is safeguards. Fable 5.2 carries cybersecurity and biology safeguards and is open to everyone, while Mythos 5.2 is limited to vetted organizations in those fields. Emergent offers Fable 5.2.

For most builds, the safeguards won't come up. If your app involves penetration testing, exploit generation, or research biology, expect Fable 5.2 to decline those requests.

When to pick Fable 5.2 over Opus 5.5

For users who don't build with Fable yet, the question is whether to step up from Opus 5.5. Pick Fable 5.2 for builds where the hard part is the thinking: complex business logic, many connected systems, or a large refactor of an existing app. Cost per task can be closer than list prices suggest: at High effort, Fable 5.2 matches Opus 5.5's best FrontierCode score at roughly half the cost per task. List pricing still reflects its top-tier position. Fable 5.2 lists at $8 per million input tokens and $40 per million output tokens on Anthropic's API, double Opus 5.5 and four times Sonnet 5.5.

Opus 5.5 remains the right default for most projects. Sonnet 5.5 still fits fast, everyday work like UI tweaks, bug fixes, and new forms.

Upgrade your hardest builds to Fable 5.2

Claude Fable 5.2 is in the Emergent model picker now. If you build with Fable 5.1, select 5.2 at the start of your next project for faster output and stronger results on long, multi-step work. Keep Opus 5.5 or Sonnet 5.5 for everything else.

Start Building on Emergent.