Google’s newly released Gemini 4 Nano is seriously impressive.

I feel like for my tasks, it can genuinely replace Opus 5.5!

Its multimodal capabilities are especially strong, as is its ability to recreate designs and animations.

This video shows a few components it just made for me, expanding on a reference image.

Because its multimodality is the real deal, it’s much better than Opus 5.5 and Claude at recreating dynamic content.

Its design style and taste are also on point, and chatting with it feels really comfortable. By comparison, GPT is absolutely useless.

Google’s only bottleneck right now is Antigravity, but it really has improved a lot lately.