OpenAI Launches GPT-5.5: Smarter, Faster, and Built for Agentic Work
OpenAI's GPT-5.5, codenamed "Spud," brings 82.7% Terminal-Bench 2.0 performance and autonomous multi-step task execution — a step toward OpenAI's unified AI super app.
OpenAI released GPT-5.5 on April 23, 2026, describing it as its "smartest and most intuitive" model to date. Internally codenamed "Spud," the release arrives one week after Anthropic's latest model update and directly takes aim at the enterprise market through agentic capabilities that go well beyond conversational AI.
Unlike its predecessors, GPT-5.5 is explicitly designed for autonomous task execution. The model can plan multi-step workflows, conduct web research, analyze large datasets, write and debug code, generate documents, and navigate computer interfaces — looping through each step independently until a defined goal is achieved. On Terminal-Bench 2.0, which measures real-world agentic coding performance, GPT-5.5 scores 82.7%, the highest recorded score for any model to date.
OpenAI co-founder Greg Brockman described it as "a faster, sharper thinker for fewer tokens," signaling that the model matches GPT-5.4's latency profile while delivering substantially higher intelligence. The model outperforms both Google Gemini 3.1 Pro and Anthropic Claude Opus 4.5 across benchmarks spanning scientific research, mathematics, and drug discovery assistance.
Availability is broad from day one: GPT-5.5 is rolling out to ChatGPT Plus, Pro, Business, and Enterprise subscribers, with a GPT-5.5 Pro variant reserved for top-tier plans. OpenAI has also integrated the model into Codex, its AI coding tool running on NVIDIA infrastructure, extending agentic capabilities to software development teams at scale.
The release is framed as a meaningful step toward a planned "super app" — a unified platform combining ChatGPT, Codex, and an AI browser under one interface. For enterprises managing fragmented AI tooling across coding assistants, research tools, and document systems, the convergence strategy could be significant. With DeepSeek V4 arriving the same day at a fraction of the closed-model price, the competition for the AI enterprise market is entering a new phase.
