Somewhat reassured, I used the fully HA-integrated Claude Code to finish migrating most of my nearly 200 smart home devices.
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.
Four chatbots agree on which team will win. I also got predictions on the coin toss, the halftime show, the most expensive ads, the national anthem and viewership numbers.