Updated
Updated · ZDNet · Aug 13
Survey Finds 75% of 138 Developers Prefer Claude Code Over Codex
Updated
Updated · ZDNet · Aug 13

Survey Finds 75% of 138 Developers Prefer Claude Code Over Codex

2 articles · Updated · ZDNet · Aug 13

Summary

  • Three in four respondents in ZDNET’s 138-developer survey said they use Claude Code, while a little more than a third use Codex and 22% use both.
  • Claude Code’s edge came mainly from handling large codebases, planning multi-file changes before editing, and delivering stronger code quality, agentic terminal workflows, and clearer reasoning.
  • Codex still won support on cost and workflow: developers cited more usage for the same money, no weekly limits, tighter ChatGPT integration, and more predictable edits with less cleanup.
  • Twenty-two respondents said they deliberately pair the tools—one writing, the other reviewing—while 14 said human review of AI-generated code remains non-negotiable.
  • The report says the poll was informal and self-selected, but it suggests developers weigh trust, pricing, and workflow almost as heavily as raw coding performance.

Insights

Why are developers abandoning cheaper AI coding tools for a terminal-based agent that demands constant supervision?
Now that AI writes the code, are human developers becoming the biggest bottleneck in software engineering?
Could your team's favorite autonomous coding agent be a silent security risk waiting to break production?