Just a few months ago, Claude Code was the absolute peak for AI programming.The tables have turned recently though, and a newcomer has taken center stage as the best LLM for programmers: Codex.Claude Code was the pinnacle of agentic coding for a while Anthropic was focusing on enterprise, and it worked well for them If you were to rewind the clock and look at the end of 2025 and beginning of 2026, Claude Code was solidly in first place when it came to agentic coding.
Google's Antigravity had gained a lot of traction, but it still lagged quite a bit behind Claude Code.You see, Claude Code is the frontier platform used by many large companies, such as Spotify.With this type of backing, Anthropic was getting quite a bit of money from API token payouts from big companies.
I have a good friend who worked for a company that relied heavily on Claude Code, and the amount of money spent on tokens was astronomical.However, Claude took its position as market leader for granted, and got a bit complacent.While Opus 4.5 was absolutely the best coding model by far at the time, and Opus 4.6 should have out performed it, Anthropic made some questionable choices.
After Opus 4.6 launched, the quality of output seemed to be drastically lower than what people expected.This, among quite a lot of downtime, pushed people to find a different solution.That solution ended up coming from OpenAI as they refreshed their Codex platform.
OpenAI's Codex quickly overtook Claude Code once it hit the shelves Where OpenAI lagged behind, now they lead the pack It's no surprise that OpenAI's ChatGPT is the most popular AI chatbot around.Anthropic's Claude is a fantastic tool, but Anthropic just doesn't have the same user reach that OpenAI does.This is why it was a bit surprising that OpenAI didn't have a solid Claude Code competitor sooner.
Sure, Codex existed as a model to be used, and it was kind of like Claude Code, but it wasn't anywhere near the level of what Anthropic was putting out.However, that changed in February 2026.When OpenAI released their brand-new Codex app, they quickly took the AI programming world by storm.
it was actually pretty surprising just how fast it took over.In a time when Claude Code was taking more tokens away from users, Codex was giving more tokens to users.In fact, the Codex rate limits reset so frequently, there's a website to track when the last time they reset was.
When Claude Code's reasoning was being turned down, Codex was offering things like fast mode and higher reasoning efforts.I was a huge Claude Code fan at first, and it was the only model I trusted for what I was working on.After a while though, I realized that Claude Code started to not live up to what it was a just a month or two prior, and I too jumped ship to the Codex train.
After using Codex for just a few days, I realized that it was definitely where I needed to spend my programming time.The output was better, the quality was high, and the features were quickly catching up to where Claude Code was.Then, Codex surpassed Claude Code in features and has kept on sailing since then.
Also, OpenAI acquired OpenClaw a few months ago shortly after it was renamed when Anthropic sent a cease and desist letter to the developer, and then started blocking people from using it with Claude models in some cases, and charging extra in others.Codex, on the other hand, works natively with OpenClaw now and allows you to use your Codex tokens from your monthly subscription with both it and Hermes.OpenAI is just being way more open with how you can use its models and your subscription, which is something that Anthropic seems to be increasingly more stingy with.
Related I'm not a programmer—but here’s why I prefer Codex over ChatGPT Despite its name, Codex isn’t just for coders and programmers—anyone can and should use it.Posts 1 By Dibakar Ghosh Codex is improving at a rate Claude Code simply can't keep up with I wanted Anthropic to succeed, but it doesn't seem likely anymore In just the past week or two, Codex has introduced a number of features, including remote coding.This means you can sync your Codex conversations with the mobile app (or other computers) and continue where you left off from anywhere.
Say you have a home Mac mini that you use for development, and an iPad for on-the-go work.You can link the iPad to your Mac mini and then continue development from anywhere without even having to set up a VPN.Just issue a command to Codex through the mobile app, and it relays it to your Mac mini.
Likewise, if you have a Mac Studio and MacBook Air, you can do the same.The MacBook Air can become a remote for the Mac Studio and you can have it issue commands there.Or, if you have a remote development box, then you can do the same type of connection just over SSH (or in the mobile app still).
Add to that computer use, plugins, and the new /goal function and Codex is simply staying ahead of Claude Code when it comes to usable features.I once was a diehard Claude Code fan, and now I don't think there's anything that could make me go back.ChatGPT+ What's included? Unlimited conversations, faster response speed, priority access, and more Brand ChatGPT ChatGPT's AI-supported assistance gets even better with a paid subscription; it Plus tier offers enhanced features including unlimited conversations, faster response speed, priority access, and more.
Try for Free Expand Collapse Codex is simply the best AI programming model out right now I now rely exclusively on Codex for my AI programming needs.It works better than Claude Code for my needs, gives me more tokens then I could ever reasonably use on the $100 per month plan, and is way more reliable than Claude Code has been over the past few months.If you're still using Claude Code, then it's time to give Codex a try.
Read More
