On Thursday, February 5th, OpenAI and Anthropic simultaneously released the latest versions of their coding-focused models: GPT-5.3-Codex and Claude Opus 4.6. Prior to this, Anthropic had captured the majority of mindshare as the industry collectively wrestled with the emerging era of agents, largely fueled by the dramatic performance leap delivered by Claude Code powered by Opus 4.5. This post doesn’t delve into how software is changing forever, how Moltbook is showcasing the future, how ML research is accelerating, or the many broader implications. Instead, it focuses on how to assess, live with, and prepare for new models. This year’s model releases will highlight the narrow performance gap between Opus 4.6 and Codex 5.3, with Opus currently leading on usability. Prior to these releases, I’d been using Claude Code heavily as a general-purpose computer agent—for software engineering as well as a wide range of data analysis, automation, and similar tasks. I’ve experimented with Codex 5.2 (typically on xhigh with maximum thinking effort), but it didn’t quite suit my wide range of horizontal tasks. Over the past few days, I’ve been using both models far more equally. As a sincere compliment, I’d say Codex 5.3 feels far more like Claude now—it’s much quicker to respond and significantly more capable across a wide range of tasks, from Git operations to data analysis. (Earlier versions, even up to 204, often struggled with basic Git commands like creating a new branch.) Codex 5.3 represents a significant advance into Claude’s domain by delivering substantially stronger product-market fit.
Interconnects AI