Feng Huang Wang·2025-08-27 03:13Core Insights - Google DeepMind has launched its advanced image generation and editing model, Gemini 2.5 Flash Image, which enhances image modification accuracy while maintaining the appearance of people and animals [1] - The new model outperforms the previous native image generation tools and shows higher accuracy in image editing tasks compared to ChatGPT's GPT-4o [1] - Gemini 2.5 Flash Image allows for precise local edits through text prompts, enabling users to blur backgrounds, remove blemishes, add colors, or erase entire objects without manual selection [1] - The model supports the integration of up to three images at once and is accessible via the Gemini App and API, with API pricing set at $30 per million output tokens and approximately $0.039 per image [1]