I've been using Gemini 2.5 Pro with Roo-Code a lot these past few days. It has really helped me a lot. I managed to get it to implemented entire features. (With some manual cleaning up at the end)

The fact that it's free for now (I know they use it for training, that's OK) is a big plus, because I've had to restart a task from scratch quite a few time. If I calculate what this would have cost me using Claude, it would have been 200-300 euros.

I've noticed that as soon as it makes a mistake (messing up the diff format is a classic), the current task is basically a total loss. For some reason, most coding tools basically just inform the model they made a mistake and should try again... but at that point, it's broken response is part of the history, and it's basically multi-shotting itself into making more mistakes. They should really just filter these out.