Holy shit, if you do LLM coding, try Claude Code, it’s a whole step beyond aider and cursor.
Discussion
Obviously.
This was my impression too after I tried it for the first time last weekend. My comparison is with ChatGPT, Llama 3.1 (biggest model) and DeepSeek
I've been using it for a few days. It's really good
I'm using Claude via goose, should be similar. It's very good.
How expensive are the credits for personal usage?
Is there a way to get it free (Like through GitHub copilot or anything) that you know of?
How does Claude Code compare with Cursor Composer in agent mode when both are using the same model? (Claude-3.7-sonnet-thinking)
I’ve been using it with Composer, and, yeah, mind-blowing. I wouldn’t actually expect the agent piping itself be much different?
Also, it’s very interesting to see how Composer usability changes when you switch between models. Claude 3.7 thinking it’s magic, with OpenAI models it’s… Not that useful, let’s say