Nano Banana 2: Google's latest AI image generation model

Fast new AI art lands, but fans want video, answers, and fewer weird refusals

TLDR: Google launched Nano Banana 2, a faster image generator with pro-level features and stronger AI labels. The crowd is excited but pushing for video generation, calling out odd refusals and a demo glitch, and debating whether this is a true upgrade or just a distilled refresh.

Google just peeled back Nano Banana 2—an AI image model promising Pro-level brains at Flash speed—rolling into Gemini, Search, and even Ads, plus stronger “made-by-AI” labels via C2PA. Sounds delicious…but the comments? They’re a whole fruit salad. The loudest chorus: “Cool stills—now make them move.” One fan begged to see images “form into videos,” saying speed means nothing if it can’t follow a script across frames. Power users are split: some cheer faster edits and better “keep the character the same” control, while others report sassy refusals about “conflicting instructions,” wondering if stricter guardrails are stepping on creativity. Drama alert: a few spotted glitches—one called out an “Umbriel” narration going silent mid-demo—fueling the “did they rush this?” gossip. Meanwhile, the tinkerers ask the spicy question: is this just a distilled version of Pro? One theory dominating the thread: Google’s got an internal pipeline that auto-spins new image models whenever a bigger base model drops—efficient if true, but also the start of endless mini-upgrades. Meme corner: banana puns everywhere (“Pro… now peel-speed”), with skeptics asking if the biggest winner is Google Ads. Verdict from the crowd: hype is real, but they want receipts—video, fewer hiccups, and proof it’s more than a rebrand.

Key Points

  • Google DeepMind launched Nano Banana 2, an AI image generation model combining Nano Banana Pro’s capabilities with Gemini Flash-level speed.
  • Nano Banana 2 is also referred to as Gemini 3.1 Flash Image.
  • The model emphasizes high-quality, photorealistic output with subject consistency and precise instruction following.
  • It is rolling out across Google products including the Gemini app, Google Search, and Ads, enabling faster editing and iteration.
  • Google is enhancing SynthID and integrating C2PA Content Credentials to help identify AI-generated content.

Hottest takes

“see how these images are starting to form into videos” — aliljet
“contains conflicting instructions” — sync
“New frontier model -> automatic new image model” — vessenes
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.