OpenAI Releases GPT-5.5, Pointing Toward an AI Super App

OpenAI released GPT-5.5 on April 23, the latest in its accelerating model family and the most direct signal yet of the company's ambition to build a unified AI super app. The new model can write and debug code, research the web, analyze data, draft documents, and operate software, planning its own path through multi-step work and checking its results as it goes.

What GPT-5.5 does differently

More than previous versions, GPT-5.5 is designed to string tasks together. Earlier models handled individual jobs well. This one is built for messy, multi-part work that involves switching between tools and navigating ambiguity without user intervention. OpenAI describes it as the foundation for what they call a super app: a single product combining ChatGPT, the Codex developer platform, and an AI-powered browser into one service.

The model supports a 1-million-token context window, letting it hold or process long documents, codebases, or conversations in one session without losing track.

Pricing and availability

GPT-5.5 is available to ChatGPT Plus, Pro, Business, and Enterprise subscribers, and through the API via Codex for developers. API pricing is $5 per million input tokens and $30 per million output tokens, roughly double the cost of GPT-5.4. OpenAI positions the higher price as reflective of the capability jump.

Performance claims

OpenAI says GPT-5.5 outperforms Google's Gemini 3.1 Pro and Anthropic's Claude Opus 4.5 on its internal evaluations. These figures are worth treating with some skepticism: AI companies select benchmarks that favor their models, and independent evaluation takes time. Competitors typically publish responses within days.

For context on where the model family now stands: GPT-5.4, released in March, was the first general-purpose OpenAI model to score above the human baseline on OSWorld, a benchmark for desktop application control, at 75.0% against a human baseline of 72.4%. It also matched or exceeded professional human performance in 83% of comparisons on GDPval, a benchmark measuring knowledge-work tasks. GPT-5.5 builds on those foundations.

The pace of releases

GPT-5.5 arrives less than two months after GPT-5.4. Analysts have started describing the current moment as a weekly sprint among frontier AI labs, with OpenAI, Google, and Anthropic pushing capable model updates faster than the industry can fully evaluate them. Whether any of these consolidates into a genuinely dominant super app remains an open question. GPT-5.5 is OpenAI's most direct answer so far.

Full details on OpenAI's announcement page. TechCrunch has more on the super app strategy.