Based on information from TestingCatalog, OpenAI is gearing up to roll out its next - generation image generation models, namely Image - 2 and Image - 2 - mini. These are set to take the place of the current Image - 1 model. Internally, during the development process, they were codenamed Chestnut and Huzzlenut respectively. The new models bring about substantial enhancements in terms of image detail and color precision. They effectively tackle the problem of a yellowish tint that was evident in the previous - generation model. When it comes to visual performance, they are on par with Google Nano Banana 2. There are remarkable improvements in how they reproduce structures and colors. These new models will amalgamate the strengths of the DALL - E series. Their target audience includes ChatGPT users as well as those utilizing OpenAI services that come with integrated image generation features. It is anticipated that they will be launched simultaneously with GPT - 5.2. This move is expected to significantly bolster OpenAI's standing and competitiveness in the realm of multimodal AI.