OpenAI has unveiled GPT-5-Codex, an advanced version of its AI coding agent integrated into Codex, boasting improved dynamic “thinking” capabilities that allow it to allocate anywhere from a few seconds to seven hours on coding tasks. This enhancement enables GPT-5-Codex to outshine its predecessor, GPT-5, in agentic coding benchmarks and code refactoring tasks, marking a significant leap in AI-assisted coding efficiency and effectiveness. The model is now available to users across various tiers, including ChatGPT Plus, Pro, Business, Edu, and Enterprise, with plans to extend access to API customers in the future.
GPT-5-Codex’s introduction is a strategic move by OpenAI to stay competitive in the bustling AI coding tool market, which has seen exponential growth and demand, leading to significant achievements like Cursor’s surpassing $500 million in ARR. The model’s ability to dynamically adjust its problem-solving duration in real time without a predetermined router sets it apart, offering a more flexible and potentially more insightful coding assistance experience. Furthermore, GPT-5-Codex has demonstrated proficiency in conducting code reviews, receiving praise from experienced software engineers for its high-impact comments and reduced error rate. This development underscores OpenAI’s commitment to refining AI’s role in software development, promising a more intuitive and powerful tool for coders worldwide.
Read more at TechCrunch…