For what it's worth, I'm not affiliated with Open AI (you can verify by my comment history [1] and account age) and I agree with the top comment. I do Elixir consulting primarily and nothing beats OpenAI's model at the moment for Elixir. Previously, their O3 models were quite decent. But, GPT-5 is really damn good.
Claude code will unnecessarily try to complicate a problem solution.
This is hilarious because for me Cursor with GPT-5 often generates Elixir that isn't even syntactically correct. It needs to be told not to use return statements, and not to try to index linked lists as arrays. Code is painfully non-idiomatic to the point of being borderline useless even in the simpler cases. Claude Sonnet 4.5 is marginally better, but not by much. Any ambitious overhaul, refactoring or large feature ends in tears and regret.
Neither tool is worth paying even $20 a month for when it comes to Elixir, that's how little value I get out of them, and it's not because I can't afford it.
Gemini is also good, I recommend you try it as well. Usually my workflow is GPT-5 as the primary, but yes, as you mentioned it is not perfect. But Gemini surprisingly compliments GPT-5 for my use cases atleast. It's good at LiveView related stuff, whereas GPT-5 is more of architecting side.
Both LLMs suck if you let it do everything without architecting the solution first. So, I always instruct the high level architecture of how I want something, specifically around how the data should flow and be consumed and what I really want to avoid. With these constraints and bit of some prompt engineering, they are actually quite good.
> Both LLMs suck if you let it do everything without architecting the solution first.
I always do that. Last time I spent an hour planning, going through the requirements, having it ask questions, only for it to completely botch the implementation.
Sure, I can treat it like a junior and spend 2-3 hours planning everything down to the individual function level and it's going to implement it alright. The code will work but it won't be idiomatic. Or I can just do it myself in 3 hours total to a much higher standard of quality, without gambling on a successful outcome, while simultaneously improving my own knowledge, understanding, and abilities.
No matter how I try to use them, agentic coding is always a net negative on my productivity (disposable one-off scripts excluded).
Personally I found Claude to be relatively OK at Elixir. With a lot of hand holding. My main problem when it comes to Elixir and Erlang is many amount of files. For that kind of boilerplate, it is good. Otherwise just use "erlang-skels.el" with Emacs. :D
[1] https://news.ycombinator.com/item?id=45491842