OpenAI launches new agentic coding model only minutes after Anthropic drops its own

On Monday, OpenAI introduced Codex, an innovative coding tool specifically designed for software developers. In a significant development, the company has also rolled out a powerful new model to enhance Codex: GPT-5.3 Codex. This updated version promises to elevate Codex from a simple coding assistant to a versatile tool capable of performing almost any task that developers and professionals undertake on a computer, thereby broadening the scope of software creation and workflow efficiency. OpenAI claims that their advanced model has been rigorously tested against various performance benchmarks, demonstrating its ability to develop intricate games and applications from inception in just a few days. Notably, GPT-5.3 Codex boasts a 25% improvement in speed compared to its predecessor, GPT-5.2. Additionally, this marks a milestone for OpenAI, as it is the first model that effectively contributed to its own development—staff utilized earlier iterations of the program to troubleshoot and assess its functionality. The unveiling of this new model follows closely after Anthropic released a competing agentic coding model. Interestingly, both companies had initially scheduled their launches for the same time—10 a.m. PST. However, Anthropic adjusted its timeline, launching 15 minutes earlier, thereby gaining a slight advantage in the race for attention in the tech community.