Lately you’ve probably seen 🍌Nano-Banana🍌 all over!
Crazy good at text-guided edits — strong prompt fidelity + scene consistency (light, angle, context).
I tested it out and, yes, it’s impressive — but honestly, with a bit of iteration you can get similar results using ChatGPT or Runway’s multi-reference 🧵
Made a whole series of absurd products with @lovart_ai, all based on historical + cultural figures 📜
Prompt’s in ALT 👀 https://t.co/OeAlQ2Rviq
Asked Midjourney to make the character images with just a super basic prompt: "PHOTOREALISTIC PORTRAIT OF [character]" 📸
I requested in @lmarena_ai to blend both images and then brought them to life with Kling AI 🎥 https://t.co/Un5k2LXneZ
@lmarena_ai That said… it’s far from perfect! I’ve had plenty of bad results, and it’s fairly slow.
Plus, each prompt only generates a single image, which makes it hard to compare with other models.
Yeah, it’s the shiny new thing — but not the best!⚡
https://t.co/pvKdSgVWj3
