OpenAI Launches GPT-5.5 to Power Super App
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI introduced GPT-5.5, an AI model built to complete complex work with little direction, according to Bloomberg's Rachel Metz.
- GPT-5.5 will power OpenAI's upcoming agentic "super app," per The Deep View, positioning the model as the backbone of the company's consumer push.
- The Information reported the model was developed under the internal codename "Spud."
- NVIDIA said it is already deploying GPT-5.5 to power Codex on its infrastructure, underscoring the deep hardware-tie-in.
- Implicator.ai framed the launch as OpenAI's bid to "reclaim enterprise lead from Anthropic," while Tae Kim's Key Context flagged a shift to a faster model release pace.
- Sam Altman outlined an "iterative deployment" strategy, saying GPT-5.5 is "already a smart model" but that the company "expect[s] rapid improvements" and treats iterative shipping as a core safety practice.
- Ethan Mollick (One Useful Thing), who had early access, called the model "very good, especially the Pro version," and posted that researchers had 5.5 run overnight experiment sweeps from only high-level ideas.
Why it matters: OpenAI is using GPT-5.5 as the connective tissue across three fronts — a consumer super app, NVIDIA-backed Codex agents, and an accelerated release cadence — putting direct pressure on Anthropic in the enterprise race while shortening the time competitors have to benchmark each new drop. The "iterative deployment" stance from Altman, echoed across coverage, signals the model cycle is now a continuous one rather than quarterly.
