Opus VS Codex...
Matthew Berman·2026-02-09 22:04
OpenAI and Anthropic just dropped their new frontier coding models, and they could not be more different. Here's what it's like to give each of them a coding task. Opus 4.6% says, "Cool. Let me go build that for you. I'll only edit the files that make the most sense for exactly the task at hand. I'll be back in a few minutes." GPT 5.3% codeex.Wow. Let me look up the entire encyclopedic knowledge of the entire universe. Then let me touch every single file that you've ever touched in your lifetime as a develo ...