ChatGPT Images 2.0 adds web access and reasoning
SkimNews Take
AI image generation is evolving beyond pure creation to integrate informational accuracy, shifting its utility toward visual communication tools rather than just artistic expression.
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- ChatGPT Images 2.0 went live in the app in both a standard version open to all users and a "thinking" mode with built-in reasoning reserved for paid subscribers.
- The new engine supports a wide range of aspect ratios and adds web access and improved typography to ChatGPT's image model.
- In hands-on testing, ChatGPT crafted a memorial image of a friend's recently deceased cat alongside two favorite toys, described as resembling a personalized sympathy card.
- The model composited two wedding photos into an old-style photo album with photo corners and turned photos of the author and her 13-year-old playing sports into trading cards complete with names, positions, and team logos extracted from their uniforms.
- Asked to build a faux newspaper called the "Smart Brevity Times" using current Axios headlines, ChatGPT initially pulled from old articles; a second attempt surfaced today's stories but looked more like a mock-up than a finished paper.
- The added reasoning means image generation takes considerably longer, and some outputs — like a mahjong cheat sheet — were accurate but lacked polish.
Why it matters: Prior AI image tools have drawn consumer interest but failed on reliability for business workflows; ChatGPT Images 2.0's reasoning mode and web access target that gap. The hands-on results suggest the model is ready for personal creative use — trading cards, memorials, collages — but still fumbles on time-sensitive sourcing and finished professional deliverables, meaning adoption will likely start with casual work before production pipelines trust it.
