Google released Gemini 4 Hydrogen on November 12. It's pitched as the fast, cheaper model for everyday coding and agent work, with a 1M-token context window, a new xhigh thinking level and a free tier in the Gemini API. It's been near the top of Hacker News since launch day.
And honestly, my first reaction wasn't excitement. It was: do I really switch again?
Every few weeks there's a new "best model for coding." Benchmarks go up, but my day-to-day work barely changes, because what decides which model stays open in my editor is never the leaderboard. It's smaller stuff:
- Does it make focused edits, or rewrite half the file I didn't ask it to touch?
- Is it fast enough that I don't switch tabs while it thinks?
- Does it say when it's unsure, instead of confidently inventing an API?
- What does it cost once the free credits run out?
I haven't run Hydrogen on real client work yet, so I don't have a verdict. That's exactly why I'm asking.
So, which model do you actually code with every day?
- What do you use it for: autocomplete, debugging, refactors, writing tests, agents?
- What made you stick with it, or what made you switch away from the last one?
- Has anyone tried Hydrogen yet? Does it feel different from what you were using?
One-line answers are welcome. I'm especially curious whether people pick one model for everything or keep a different one for different jobs.
Top comments (0)