How Cursor uses GPT-5

Source: OpenAI Date: 2025-08-07 URL: https://openai.com/index/gpt-5-cursor

Summary

Summary

A case study covering Cursor’s use of GPT-5 in its AI-native code editor — describing how GPT-5’s improved instruction-following, longer context window, and reasoning capabilities enhanced Cursor’s code completion, explanation, and refactoring features. Cursor had become one of the highest-profile OpenAI API customers, and their GPT-5 adoption reflected the rapid improvement in coding AI experiences.

Implications

Platform/agentic engineering thread. Cursor’s GPT-5 case study is significant because Cursor competes with GitHub Copilot — meaning it represents a case where an OpenAI API customer built a better coding experience than Microsoft’s own product. The case study is OpenAI demonstrating that the API model is competitive with the integrated product model for developer tools. GPT-5’s improvements (stronger instruction-following, better context utilization, deeper code understanding) translated directly to Cursor’s product quality, illustrating the flywheel: better foundation models → better Cursor → more developer adoption → more API revenue. The August 2025 timing positions this alongside GPT-5’s broader professional capability rollout.