Search
Cursor Enhances Code Completion with Tab Model Upgrade
2025-09-15 /
Read about 0 minute
Author：小编
As reported by analyticsindiamag, the AI-powered coding platform Cursor has unveiled a major enhancement to its code autocompletion system, known as the Tab model. This upgrade is specifically designed to minimize the occurrence of low-quality suggestions and to boost the overall accuracy of the system. Cursor utilizes a reinforcement learning strategy based on a gradient approach. In this setup, the model is rewarded when its suggestions are accepted by users, while it incurs penalties for suggestions that are rejected. At present, the Tab model is handling an impressive volume of over 400 million requests on a daily basis.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
