On Monday, OpenAI unveiled Codex, an advanced agentic coding tool designed for software developers. Just a day later, the company launched GPT-5.3 Codex, a significant upgrade to its existing system. OpenAI claims the new model transforms Codex from a tool that can only write and review code into one capable of performing nearly any task a developer or professional might do on a computer. This evolution, the company says, has the potential to democratize software creation and reshape how work is carried out. According to OpenAI, GPT-5.3 Codex can generate fully functional complex games and applications from scratch within days. The model is also reported to be 25% faster than its predecessor, GPT-5.2. One of its most notable features is that it was the first model developed by OpenAI to be instrumental in its own creation—early versions were used by the company’s engineers to debug and evaluate performance, marking a milestone in self-improving AI systems. The release comes in direct response to a similar announcement from competitor Anthropic, which launched its own agentic coding model just minutes before OpenAI’s updated version. Both companies had initially planned to roll out their tools simultaneously at 10 a.m. PST. However, Anthropic shifted its release time up by 15 minutes, narrowly beating OpenAI to the public spotlight. The rapid back-and-forth between the two AI leaders underscores the intense competition in the race to develop next-generation tools that can autonomously handle complex, real-world programming tasks. With both models designed to act as intelligent agents capable of end-to-end software development, the release marks a pivotal moment in the evolution of AI-powered coding assistants.