OpenAI lifts baseline quality
- OpenAI rolled GPT-5.4 ‘Thinking’ into ChatGPT, targeting harder reasoning tasks like spreadsheets, code, and document understanding. - It also released ChatGPT Images 2.0 with stronger realism, better aspect ratios, and reasoning-driven image generation. - The upgrades raise the baseline for acceptable AI output, increasing expectations for human editing and design judgment (help.openai.com) (dataconomy.com)
OpenAI has folded a stronger reasoning model and a new image generator into ChatGPT, raising the default quality users now get out of the box. (help.openai.com) OpenAI’s help center says GPT-5.4 Thinking launched in ChatGPT in March 2026, with a focus on coding, spreadsheets, presentations, and document-heavy work that usually takes more back-and-forth. (help.openai.com) (openai.com) OpenAI also said GPT-5.4 Thinking can show an upfront plan while it works, and the company’s system card compares it mainly against GPT-5.2 Thinking because there is no GPT-5.3 Thinking model. (openai.com 1) (openai.com 2) On April 22, 2026, OpenAI’s ChatGPT release notes said ChatGPT Images 2.0 rolled out to all plans, while “images with thinking” became available on paid plans when users select Thinking and Pro models. (help.openai.com) The image update changes how ChatGPT makes pictures. Instead of treating image generation as a one-shot prompt, OpenAI says the model can spend extra time planning and refining before it renders the final image. (help.openai.com) (dataconomy.com) That shift lands after a year in which ChatGPT’s image tools moved from novelty to routine work. OpenAI’s image help page says users can now generate and edit images directly in chat, including selecting part of an image and asking for a targeted change. (help.openai.com) OpenAI has also been pruning older ChatGPT models as it adds newer ones. Its help center says GPT-4o, GPT-4.1, GPT-4.1 mini, OpenAI o4-mini, and GPT-5 Instant and Thinking were retired from ChatGPT on February 13, 2026, though API access was unchanged. (help.openai.com) The practical effect is that users now start from a higher floor. If ChatGPT can produce cleaner code, stronger document analysis, and more polished visuals on the first pass, the work shifts toward checking facts, tightening language, and deciding what should be kept. (openai.com) (help.openai.com) OpenAI’s own product notes frame both releases as tools for professional tasks rather than demos: harder reasoning on one side, more controllable visuals on the other. In ChatGPT, that means the baseline answer now arrives closer to finished work than rough draft. (openai.com) (help.openai.com)
