Claude Fable 5 Is Back in Cursor
Claude Fable 5 has returned to Cursor and tops CursorBench across all models — though it comes at a higher cost per task than its alternatives.
After a brief absence, Claude Fable 5 is available again in Cursor — and it holds the top spot on CursorBench, outperforming every other model available in the editor.
Fable 5 Returns to the Top of CursorBench
Claude Fable 5 leads all models on CursorBench, Cursor's own internal benchmark for measuring how well a model performs real coding tasks inside the editor. That makes it the highest-capability option currently available for Cursor users — particularly suited to complex, multi-step coding work where accuracy and agentic reasoning matter most.
The Trade-off: Top Performance at a Higher Cost
The catch is price. Fable 5 is the most expensive model per task in Cursor's lineup. Developers who need maximum accuracy on demanding tasks will find it well worth it, but those working on lighter edits or operating under tighter credit budgets may prefer sticking with a faster, lower-cost option like Sonnet 5 for everyday work.
Why It Matters
For creators and developers who rely on Cursor for serious coding projects, having Fable 5 back means access to the strongest reasoning Anthropic offers in an agentic code editor context. Whether you're scaffolding a complex feature, debugging tricky logic, or running long autonomous sessions, Fable 5 is now your highest-ceiling option in Cursor again.
Sources: the announcement on X.