ChatGPT Images 2.0 launches
- OpenAI released ChatGPT Images 2.0, an upgraded image generation and editing tool promoted as having 'thinking-level intelligence.' (x.com) - The announcement drew major engagement: roughly 23k likes and reported views in the millions across platforms. (x.com) - Creators and critics immediately flagged new editing features and workflow implications for digital art and commercial imagery. (x.com)
OpenAI rolled out ChatGPT Images 2.0 on April 21, adding a new image model and a “thinking” mode to generate and edit pictures inside ChatGPT. (openai.com) OpenAI said the update is available on all ChatGPT plans, while “images with thinking” is available on paid plans and coming soon to Enterprise and Edu. The company also released the model to developers in the application programming interface and in Codex under the name `gpt-image-2`. (help.openai.com) (community.openai.com) The new model is built to handle text-heavy visuals, multilingual layouts, and more exact edits, according to OpenAI’s launch materials. OpenAI said it supports more aspect ratios and resolutions up to 2K in the application programming interface. (openai.com) (community.openai.com) Image generation systems turn text prompts into pictures, and editing tools let users upload an image and change part of it with a selection tool or a text instruction. OpenAI’s help documentation says ChatGPT Images can add text, remove elements, make backgrounds transparent, and revise only selected areas of an image. (help.openai.com) OpenAI’s release notes say the “thinking” option gives the system more time to plan and refine an image before it is generated. In the developer announcement, OpenAI said reasoning models can research, transform inputs, generate variations, and self-check before producing an asset. (help.openai.com) (community.openai.com) The launch lands four months after OpenAI introduced “the new ChatGPT Images” in December 2025, when it said the earlier model delivered more precise edits and faster generation. VentureBeat reported ChatGPT Images 2.0 as the next major step after that GPT-Image-1.5 release. (openai.com) (venturebeat.com) OpenAI’s developer post said `gpt-image-2` reached No. 1 across Image Arena leaderboards within hours of launch, including a reported 242-point lead in text-to-image. That claim came from OpenAI’s own announcement, not an independently published benchmark paper. (community.openai.com) Early coverage focused on commercial uses as much as art. VentureBeat said testers were using the model for infographics, maps, floor plans, slides, user-interface mockups, and character sheets, while PetaPixel said OpenAI was pitching the system for design, education, development, and content creation workflows. (venturebeat.com) (petapixel.com) The immediate test for ChatGPT Images 2.0 is whether those polished demos hold up in everyday use, where designers and marketers care less about novelty than readable text, consistent layouts, and edits that do not need cleanup. OpenAI’s own help page still notes that selections are “not always precise,” and edits can spill beyond the highlighted area. (help.openai.com)
