In the realm of AI programming assistants, a fierce competition is underway. Recent data reveals that OpenAI's programming assistant, Codex, has surpassed Anthropic's Claude Code in certain core capabilities. On October 9th, startup Modu collected data from over 300,000 code pull requests, showing that developers approved the code generated by Codex at a rate of 74.3%, slightly higher than the 73.7% approval rate for Claude Code.