OpenAI has unveiled GPT-5.3 Codex, an advanced agentic coding model designed to significantly enhance its existing Codex tool. This launch comes just minutes after rival AI firm Anthropic released its own agentic coding solution, highlighting the intensifying competition in the artificial intelligence development space.
The new GPT-5.3 Codex model is positioned as a major upgrade to the original Codex, which OpenAI introduced earlier this week as an agentic coding tool for software developers. According to OpenAI, GPT-5.3 Codex transforms its predecessor from a system capable of merely "write and review code" into an agent that can perform "nearly anything developers and professionals do on a computer," thereby broadening access to software creation and streamlining workflows. OpenAI asserts that, based on internal benchmarks, the model can generate "highly functional complex games and apps from scratch over the course of days."
Beyond its expanded capabilities, GPT-5.3 Codex boasts a 25% speed improvement over its previous iteration, GPT-5.2. Notably, OpenAI also claims that GPT-5.3 Codex was "instrumental in creating itself," with company staff utilizing early versions of the program for debugging and performance evaluation during its development.
This strategic release by OpenAI closely follows the introduction of a new agentic coding model by its competitor, Anthropic. Interestingly, both OpenAI and Anthropic had initially scheduled their respective agentic coding tool launches for the exact same time: 10 a.m. PST. However, Anthropic reportedly moved its release forward by 15 minutes, securing a slight lead in the public announcement of these advanced AI models.
