Base44 Adds Fable 5: Anthropic's Most Capable Public Model

Fable 5 Joins the Base44 Model Selector

Base44 added Claude Fable 5 to its AI model selector on June 10, 2026, giving builders access to Anthropic's most capable publicly available model directly inside the platform. Fable 5 is the first model drawn from Anthropic's new Mythos class, a tier that sits above the Opus family in raw capability. It arrives in Base44 alongside the existing model lineup, selectable for any project where the extra reasoning power is worth it.

What Fable 5 Brings to App Building

Fable 5 is built for long-horizon, autonomous work rather than quick single-turn answers. That focus maps directly onto the kind of work Base44 users do: scaffolding a full application, wiring up multiple connected features, and debugging issues that span many files and several steps of reasoning.

The model maintains coherent focus across very large contexts, so it holds the shape of a growing project in mind instead of losing track as the app expands. Anthropic also reports a marked improvement on vision-based tasks, including reconstructing application source code from screenshots, which is useful when turning a design or reference image into a working Base44 app. On memory-dependent tasks, Fable 5 shows a threefold improvement over earlier models, helping it carry context forward across a longer build session.

When to Reach for Fable 5

Base44 positions Fable 5 as the model to choose for complex, multi-step builds and deep debugging, rather than as a default for every prompt. For a straightforward change or a quick iteration, the lighter models remain a faster and cheaper choice. But when a build involves intricate logic, many moving parts, or a stubborn bug that has resisted earlier attempts, switching to Fable 5 brings noticeably stronger reasoning to bear.

Pricing and Availability

Fable 5 is included at no additional cost across Anthropic's Pro, Max, Team, and seat-based Enterprise subscription plans through June 22, 2026. After that date, access moves to usage-based pricing, with the underlying API priced at $10 per million input tokens and $50 per million output tokens. In a small set of high-risk categories, Fable 5 transparently falls back to Claude Opus 4.8; Anthropic reports this happens in fewer than 5% of sessions.