In my experience Claude 3.7 is far superior for coding than Gemini 2.5. I tried it in Cursor and I wanted it to work, as a recent ex-Googler. I repeatedly found it inferior. I think it’s still behind Claud 3.5 for coding.
It would decide arbitrarily not to finish tasks and suggest that I do them. It made simple errors and failed to catch them.
What's the difference between Cline and Roo Code now? Originally Roo was a fork of Cline that added a couple of extra settings. But now it seems like an entirely different app, with it's own website even.
It depends on the task, and prompting feels different.
I've found that sonnet is possibly better at starting things from scratch and frontend code, while Gemini has been able to one-shot difficult problems and fix bugs that sonnet has struggled with.
Switching between them is a frequent occurrence for me.
It might be relevant that I've completely stopped using Cursor in favor of other tools/agents.
I've also created a few "agents" to do specific tasks using Probe[0] as an MCP server, although I'm sure you could create a full-fledged agent with it if you wanted to.
Cursor is likely very tuned for Claude (prompts-wise and all) due to its dominance with 3.5 and now 3.7. Still, Gemini 2.5's tool calling has been pretty poor in my experience which Cursor heavily relies on.
Same here. I've seen some articles and LLM benchmarks that Gemini 2.5 Pro is better than Claude 3.7 on coding, but base on my recent experience of solving code problems with two products, Claude still gave me better answer, Gemini response are more detail and well structured, but less accurate.
Same. I went back from Gemini to Claude yesterday, because Gemini was writing decidedly worse code, at times not even able to stick to Python syntax. Using Aider.
It would decide arbitrarily not to finish tasks and suggest that I do them. It made simple errors and failed to catch them.