Google 刚刚发布了 Nano Banana 2,这感觉像是 AI 图像生成领域一个真正的拐点。

Google just dropped Nano Banana 2 and this feels like a genuine inflection point for AI image generation. The model pulls real time information from web search and integrates live weather data, local conditions, actual world context into image generation at 2K/4K resolution. That is not just a style upgrade. That is a fundamentally different approach to how AI understands what it creates. Rolling out today across Gemini, Search in 141 countries, Google AI Studio, Vertex AI, and Antigravity. The speed here is wild. Flash level performance with Pro level quality. The Window Seat demo alone should make every designer pause. It generates accurate views from any window on the planet using live local data. Character consistency for up to five characters, fidelity across 14 objects in a single workflow, new aspect ratios including 4:1 and 8:1. I keep wondering when OpenAI and Midjourney will respond. Google is shipping at a pace that makes the competition look asleep.

Google 刚刚发布了 Nano Banana 2,这感觉像是 AI 图像生成领域一个真正的拐点。 该模型从网络搜索中获取实时信息,并将实时天气数据、当地条件、真实世界上下文整合到 2K/4K 分辨率的图像生成中。这不仅仅是一次风格升级,而是一种根本不同的方法,改变了 AI 理解其所创建内容的方式。 今天起在 Gemini、141 个国家的 Search、Google AI Studio、Vertex AI 和 Antigravity 上推出。这里的速度简直惊人。Flash 级别的性能,Pro 级别的质量。 仅凭 Window Seat 演示就足以让每位设计师驻足思考。它利用实时本地数据,生成从地球上任何一扇窗户看出去的准确视图。单个工作流中支持最多 5 个角色的角色一致性、14 个对象的保真度,以及包括 4:1 和 8:1 在内的新宽高比。 我一直在想 OpenAI 和 Midjourney 什么时候会做出回应。Google 的发布节奏让竞争对手看起来像是在打瞌睡。