OpenAI Unveils ‘Image 2.0’ for Text-Perfect AI Graphics Renders Bengali and Other languages
One of the biggest challenges in AI-generated images has been rendering accurate text and spelling within visuals. To overcome this limitation, OpenAI has introduced a new “Image 2.0” model in ChatGPT. Unveiled on Tuesday, the new version demonstrates remarkable proficiency in generating clear and accurate text within images.
According to a report by TechCrunch, earlier AI-generated visuals—such as restaurant menus or posters—often contained distorted or nonsensical spellings. However, “Image 2.0” can now render text with a high degree of accuracy. The model incorporates advanced “thinking capabilities,” enabling it to review and refine its own generated images. As a result, creating multi-panel comic strips, marketing assets, or complex UI designs is now easier and more reliable.
For users in Bangladesh, a key highlight is that the new model can accurately render not only Latin scripts but also Bengali, Hindi, Japanese, and Korean scripts with significantly improved clarity. The 2K-resolution model can precisely follow detailed user instructions.
OpenAI stated that both free and premium users of ChatGPT will have access to this feature, although paid users will benefit from more advanced and complex output capabilities. In addition to image generation, a new API has also been released for developers. Although the model’s knowledge base is limited up to December 2025, it is already being regarded as a new milestone in the field of AI image generation.
DBTech/BMT/OR
