✳flâneur — a map of the web's best reading

Near-Instant Full-File Edits

cursor.com · 1,555 words · saved by 1 readers

Frontier models such as GPT-4o struggle on large edits, with problems of laziness, inaccuracy, and high-latency. This is a weakness visible in coding agents. Accurately editing hundreds of lines can take multiple model calls, at times trapping the agent in an infinite loop. Even small, isolated edits are plagued with bugs: Worst of all, existing models are slow at large edits, breaking the programmer out of flow. We've trained a specialized model on an important version of the full-file code edit task called fast apply. Difficult code edits can be broken down into two stages: planning, and applying. In Cursor, the planning phase takes the form of a chat interface with a powerful frontier model. Applying the change to the current file should be straightforward and instant. Our fast-apply model surpasses GPT-4 and GPT-4o performance and pushes the pareto frontier on the accuracy / latency curve. We achieve speeds of ~1000 tokens/s (around 3500 char/s) on our 70b model using a speculative

Blog / research Frontier models such as GPT-4o struggle on large edits, with problems of laziness, inaccuracy, and high-latency. This is a weakness visible in coding agents. Accurately editing hundreds of lines can take multiple model calls, at times trapping the agent in an infinite loop. Even small, isolated edits are plagued with bugs. Worst of all, existing models are slow at large edits, breaking the programmer out of flow. We've trained a specialized model on an important version of the full-file code edit task called fast apply . Difficult code edits can be broken down into two stages:

Explore this link on the map →

related reading