Hook

Their other posts in the index, biggest breakout first.
This is my updated AI coding workflow as of December 2025. Cursor is still my primary IDE, but I only pay for the $20 base subscription because I want to pay for that tab to complete model. I've tried a couple of other solutions out there, but I still find Cursor to have the best tab to complete experience. Claude Code handles the bulk of my agent to coding workflow, and I pay for the 20x max plan, which is the $200 a month plan. Super expensive, but very worth it in my opinion. And I pretty much always default to using the Opus 4.5 model. But the new addition that I recently added into the mix is a lot more Gemini 3 into my workflow. I think right now, Gemini 3 is the best model out there for UI/UX and frontend development in general. So I use it whenever I need to do any type of frontend work for the various apps that I'm building. I primarily use it within the Gemini CLI tool because I do have a Gemini Pro subscription, so I do have access to it. But sometimes I also use Gemini 3 directly within Cursor as well, which is a part of my $20 a month subscription that I pay for. I actually want to use more Gemini 3 in my day-to-day coding to replace some usage of Claude Code, but quite frankly the Gemini CLI user experience is the main thing that's holding me back. If I can use the same exact Claude Code user interface within the terminal, but then just also hit the Gemini 3 API endpoints, I'll be so happy to do that. I really do think Gemini 3 is a great model, but for the time being Claude Code Opus 4.5 takes up the vast majority of my day-to-day coding work still. That is my current AI coding setup. I'm curious to hear what yours is, lemme know in the comments down below.