Re:6 Weeks of Claude Code
by cpursley on Saturday August 02, 2025 @03:49PM
(#44770781)
Gemini is shockingly, embarrassingly, shamefully bad (for something out of a company like Google). Even the open models like Qwen and Kimi are better on opencode.
[ Parent
| on HN ]
Re:6 Weeks of Claude Code
by indigodaddy on Saturday August 02, 2025 @03:56PM
(#44770844)
Ah I was thinking maybe the Gemini-cli agent itself might be attributable to the problems, this maybe try the opencode/Gemini combo instead..
I'd like to mess around with "opencode+copilot free auth" or "{opencode|crush}+model via grow(still free?)" to see what kind of mileage I can get and if it's halfway decent..
In my experience, Gemini is pretty good in multishotting. So just give it a system prompt, some example user/assistant pairs, and it can produce great results!
And this is its biggest weakness for coding. As soon as it makes a single mistake, it's over. It somehow has learned that during this "conversation" it's having, it should make that mistake over and over again. And then it starts saying things like "Wow, I'm really messing up the diff format!"