OpenAI today officially launched an API for its latest AI model, GPT-5-Codex, which demonstrates groundbreaking capabilities in programming. The model maintains the same pricing strategy as GPT-5, charging $1.25 per million input tokens and $10 per output token. This new tool promises to revolutionize the way developers work, thanks to its powerful contextual processing capabilities and intelligent programming support.
GPT-5-Codex's core advantage lies in its large 400k context window, enabling it to handle more complex codebases and long text tasks. The knowledge cutoff has been updated to September 30, 2024, ensuring cutting-edge technical recommendations. In programming performance tests, the model outperformed GPT-5 on the SWE-bench benchmark, particularly excelling at large codebase refactoring tasks. Its "thinking" times ranged from a few seconds to seven hours, demonstrating its flexibility in solving complex problems.
OpenAI is making this technology available through an API, enabling seamless integration into various development platforms. Enterprises can leverage it to optimize their software development processes, while developers can receive real-time code suggestions and bug fixes. This innovation not only lowers the threshold for programming, but also elevates AI collaboration to a new level, marking the full arrival of the era of intelligent programming.
