Top News

Google has introduced Nano Banana Pro, a higher‑fidelity image generation and editing model built on Gemini 3 Pro Image, designed to produce more photorealistic, production‑ready visuals with tighter control over lighting, lens effects, composition, and on‑image text. The model is rolling out in a paid preview for developers and enterprises via Google AI Studio and Vertex AI, with a focus on professional workflows that need precise art direction, consistent characters, complex scene composition, and accurate typography inside generated images. The upgrade builds on the earlier Nano Banana (Gemini 2.5 Flash Image) and raises quality ceilings across realism, text rendering, and reasoning‑driven scene logic.

What’s new in Nano Banana Pro

Gemini 3 Pro Image backbone: The model inherits Gemini 3’s stronger world knowledge and reasoning, yielding scenes that obey real‑world physics (perspective, shadows, depth of field) and reduce common artefacts in hands, small faces, and fine textures.

Higher fidelity and controls: Prompts can specify lens focal length, aperture, shutter style, colour grading, and lighting setups (e.g., “softbox to camera left, rim light at 45°”), enabling studio‑style direction and repeatable looks across batches.

Sharper, longer text rendering: The system materially improves on‑image typography, handling long strings and complex letterforms for posters, interfaces, and product mockups with greater legibility.

Grounded generation: Optional grounding lets the model consult the web to align details (e.g., uniforms, skylines, product references) so outputs are more consistent with real‑world facts when requested.

Bigger canvases: Output sizes rise up to 4K‑class PNGs in the preview, expanding headroom for print‑quality assets and detailed UI comps.

Multiturn editing: Conversational edits preserve context and spatial intent, such as “move the key light warmer,” “swap the packaging to berry flavour,” “keep the same model and pose but change the backdrop to Santorini at golden hour.”

How it compares to Nano Banana (Gemini 2.5 Flash Image)

Quality and realism: Pro produces more lifelike skin, materials, and lighting, with fewer compositional glitches; Flash remains faster for iterative ideation and quick social‑scale renders.

Text in images: Pro’s typography engine is notably better for signage, UI frames, ads, and long labels, reducing the need for manual touch‑ups.

Canvas and control: Pro supports higher resolutions and more granular scene control; Flash prioritises speed and lower latency.