Hi Cursor team,
Today I came across an open source text classification model called notdiamond-0001 that automatically determines whether a query is better suited for GPT-3.5 or GPT-4:
This seems like it could be useful for Cursor to intelligently route code completion prompts between GPT-3.5 and GPT-4. By using a classifier to pick the optimal model per query, Cursor users may be able to faster, less expensive queries for easier tasks.
The notdiamond-0001 model is free to use and integrate under the Apache 2.0 license.
Thank you!