Hundreds of agents tried building a browser for a week, producing 1M lines of code, 1,000's of files yet missed web standards ...
In tests, Gemini 3.5 outpaced GPT-5.1 High and Opus 4.5 on coding tasks, giving you cleaner outputs and fewer fix-it loops.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results