New-ZZZ
RU / EN
Developer Tools 1 September 2026

Claude Fable 5.1 Sets a CursorBench Record

N
New-ZZZ desk
X @cursor_ai · 1 day ago

The Claude Fable 5.1 model is now available to Cursor users. In the CursorBench 3.2 test, it achieved the best result among the models evaluated by the company—73.4% at the maximum level of computational effort.

Cursor’s developers note that the model is particularly good at double-checking its own solutions. This helps it independently complete complex programming tasks from start to finish.

Why it matters

  • The model achieved the best result among the solutions tested by Cursor on CursorBench 3.2.
  • Its ability to double-check its own work improves the reliability of autonomously solving complex programming tasks.

Key facts

  • Claude Fable 5.1 is now available in Cursor.
  • The model scored 73.4% on CursorBench 3.2 at maximum compute effort.
  • Cursor highlights the model's ability to verify its own work.
  • The model is designed to complete complex programming tasks from start to finish.
Read the original

The full text is in the original source. Here we provide a brief summary and key facts.

/ related