Which of these do you prefer for coding ? In last few days, it seems Codex has started outperforming Claude Code

All these models are built on same underlying technology - LLMs based on autoregressive transformers. No company has any big advantage or moat over others.

So it’s a game of cat & mouse. Leaderboards will keep changing every week, as this or that model makes a small improvement due to latest tweak, and others will catch up soon.

So don’t fret about it too much. Pick a model / tool, whichever you’re comfortable with, and stick with it. In the long run, all the models will have roughly similar performance, as they catch up with one another week by week in the cat & mouse game.

Codex hasn't been out performing claude code in the last few weeks. It's just that people who are leaving claude code want to shift to a cheaper alternative - which is codex. They are hoping codex would be similar to claude , but I've read that codex is not good either.

If u want cheaper option, take 250rs subscription of glm4.5 ..and use it with claude code...it's good

When I tried Claude it did not appeal to me. It's outputs were not that refined. Codex I never used so don't know. There are better platforms out there than Claude.

Claude has been consistently outperforming

Codex for me as the org pays for it

Claude sonet is pro model

Right now claude is not working properly, they have downgraded performance, it's either model issue or claude cli issue

Wat works in open ai codex..it's too good

Aider AI

Claude by far!!