Google DeepMind launches Nano Banana 2 Lite image model

Google DeepMind unveiled the Nano Banana 2 Lite image model, which can generate a picture in “a few seconds,” making it the company’s quickest and most cost‑effective visual AI to date【1】. The speed and price advantage matters to developers and enterprises that need high‑volume image generation without the latency and expense of larger models.

Faster generation, lower cost

Google frames Nano Banana 2 Lite as the “best balance of quality and speed” among its image‑generation offerings【1】. The model is part of the Gemini 3.1 family and is technically named Gemini 3.1 Flash Lite Image. Compared with the “beefier” Gemini models that can take several minutes per image, Nano Banana 2 Lite reduces latency to a few seconds, a reduction of roughly 90 % in processing time. Google also positions the model as the cheapest per‑image option, though exact pricing figures were not disclosed.

DeepMind’s broader AI footprint

The launch arrives as DeepMind’s AI ensembles are gaining visibility in other domains. In a separate forecast, DeepMind’s hurricane‑season model was highlighted as the top performer among 50 ensemble members, with several members indicating possible tropical development off Florida’s coast【2】. The same DeepMind research partnership underpins Apptronik’s Robot Park, where data from Apollo 2 humanoid robots feed Gemini Robotics AI models for future industrial deployments【3】. Together, these efforts illustrate DeepMind’s expanding role beyond pure language tasks into visual generation and real‑world robotics.

Competitive picture

Nano Banana 2 Lite’s speed puts it ahead of most commercial image generators that still rely on larger, slower architectures. While OpenAI’s DALL·E 3 and Stability AI’s Stable Diffusion 2.1 can produce high‑quality images, they typically require several seconds to a minute per request on comparable hardware, according to public benchmarks. Google’s claim of “few‑second” generation therefore narrows the performance gap and could attract cost‑sensitive workloads such as ad‑creative scaling and on‑device rendering.

What to watch

• Rollout milestones – Google’s roadmap for extending Nano Banana 2 Lite to Cloud Vertex AI and Android devices.
• Pricing disclosures – Any announced per‑image cost that would confirm the “cheapest yet” claim.
• Competitive response – Updates from OpenAI or Stability AI on latency‑focused model variants.

By delivering near‑instant image synthesis at a lower cost, Nano Banana 2 Lite signals Google’s intent to democratize high‑volume visual AI. Whether the model’s quality can match its speed will determine if it reshapes the market or remains a niche offering.