OpenAI launched ChatGPT Images 2.0 on April 21, 2026, making the new model available to all users starting April 22. The model — backed by the gpt-image-2 API — produces images at up to 2K resolution with dramatically improved text rendering including non-Latin scripts (Japanese, Korean, Hindi, Bengali) and an optional "thinking" mode that lets the model search the web, generate multiple variants from one prompt, and verify its own outputs before delivery. Paid subscribers get thinking mode; free users get the standard version. TechCrunch notes the model handles dense text, UI elements, and multi-panel comic strips — content that reliably broke previous image models — though complex generations take several minutes rather than seconds.