实景 grounding：首款融合 Google 图片搜索 grounding 的图像模型，可检索地标、品牌、物体等真实视觉参考，保证输出准确写实。
Factual Grounding: First image model integrating grounding with Google ImageSearch, retrieving real-world visual references (landmarks, brands, objects) for accurate, reality-based outputs.
高级文字渲染：直接在图像生成清晰、准确文字，支持多语言与图内翻译。
Advanced Text Rendering: Renders sharp, accurate text directly in images, supporting multiple languages and in-image translation.
创作一致性：单流程可保持最多 5 个角色统一、14 个物体细节一致。
Creative Consistency: Maintains up to 5 consistent characters and 14 consistent objects in one workflow.
视觉质感：光影、纹理、清晰度相较 NB1 大幅提升。
Visual Fidelity: Significant upgrades in lighting, textures, and detail sharpness over NB1.
