OpenAI is set to unveil two new image generation models, Chestnut and Huzzlenut, alongside the anticipated GPT-5.2 release. These models, referred to as Image-2 and Image-2-mini, aim to enhance visual quality and versatility compared to the existing Image-1. Early testers highlight improvements in image detail and fidelity, effectively addressing past issues like the yellow tint in Image-1 outputs. These advancements bring OpenAI closer to Google’s Nano Banana 2 in rendering detail and color accuracy. While Nano Banana 2 remains superior in specific scenarios, the gap has narrowed significantly. The new visual styles echo earlier DALL-E models, integrating past strengths with modern rendering capabilities. Although an official launch date is yet to be announced, indications in ChatGPT’s code suggest a synchronized release with GPT-5.2. This initiative aligns with OpenAI’s strategy to enhance its AI offerings, benefiting creative professionals and businesses through higher-quality generated assets.
Source link
Share
Read more