Comment by sreekanth850

4 days ago

Open AI seems to be jack of all trades.i randomly use chatgpt for random questions, never for a serious task. They should check how anthropic is laserfocussed on coding and b2b segment.

5 comments

sreekanth850

Reply
moomoo11  3 days ago

Codex 5.3 is way better than opus 4.6

  • sreekanth850  3 days ago

    My experience is completely the opposite. For generic, low-complexity CRUD tasks, Codex works fine. But when it comes to complex bug fixing, it completely fails, especially with middleware pipelines and complex authentication issues. Gemini also shines, codex is absolutely terrible for complex coding.

  • christophilus  3 days ago

    I’ve used Claude Code almost exclusively since its release. In the last week or two, I gave Codex a spin. So far, I’m impressed. It does seem to have reached parity. And, it doesn’t run out of tokens nearly as fast as the equivalent Claude plan.

    • moomoo11  3 days ago

      I stopped using Claude Code because Anthropic lies about the token usage. There were multiple github issues about it, and they just ignore/close them.

      Codex 5.3 feels magical, I can have it continue in a single session for hours and it doesn't choke one bit. It is insane lol