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

On Monday, OpenAI introduced Codex, an AI-powered coding tool for software developers. Shortly afterwards, the company followed up with another major release: a new model explicitly built to supercharge Codex, called GPT-5.3 Codex.

According to OpenAI, the new model upgrades Codex from an agent that can "write and review code" into one capable of doing "nearly anything developers and professionals do on a computer, expanding who can build software and how work gets done." After testing GPT-5.3 Codex across a range of internal performance benchmarks, OpenAI says the model can build "highly functional complex games and apps from scratch over the course of days."

OpenAI also said that GPT-5.3 Codex delivers a significant speed improvement, running about 25% faster than its previous version, GPT-5.2. The company added that this is the first OpenAI model that "was instrumental in creating itself," noting that OpenAI's own engineers used early versions of GPT-5.3 Codex to debug the system and assess its performance during development.

The timing of the release is notable, as it comes immediately after Anthropic's competing announcement, which unveiled its own agentic coding model earlier that day. In fact, OpenAI and Anthropic had initially planned to release their respective agentic coding tools at the same time, 10 a.m. Pacific.

However, shortly before the scheduled launch, Anthropic moved its announcement forward by 15 minutes, narrowly beating OpenAI to market in what has become a closely watched race between the two AI labs to define the future of agentic coding tools.