OpenAI launches GPT‑5.5, boosting coding computer use
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI unveiled GPT‑5.5, a model built to handle complex tasks with limited instructions and to power its upcoming “super app.”
- GPT‑5.5 delivers its biggest improvements in agentic coding, computer use, and early scientific research, requiring reasoning across longer contexts.
- Codex in GPT‑5.5 adds an “auto‑review” mode that lets the model run longer automation tasks with fewer approvals while a separate safety agent checks high‑risk steps.
- GPT‑5.5 can see screen content, click, type, and navigate desktop applications, expanding its computer‑use abilities.
- Early testers report GPT‑5.5 writes higher‑quality code, runs research experiments overnight from high‑level ideas, and produces better spreadsheets, slide decks, and documents in Office and Google suites.
- OpenAI leadership describes GPT‑5.5 as a “new class of intelligence” that is token‑efficient, low‑latency, and scalable for large‑scale work.
Why it matters: Developers and researchers gain a more capable assistant that can write code, run experiments, and manipulate desktop apps with minimal prompting, while OpenAI secures a competitive edge by embedding the model in its upcoming super‑app, potentially reshaping enterprise AI workflows.
