OpenAI Unveils GPT-5.5: Massive Cost Cuts and Performance Leap with NVIDIA

OpenAI has released GPT-5.5, codenamed 'Spud,' on April 23, 2026, with API access following the next day. The model sets new benchmarks in AI performance and operational efficiency. It scored 82.7% on Terminal-Bench 2.0 and 51.7% on FrontierMath, outperforming competitors like Claude Opus 4.7 and Gemini 3.1 Pro. Built on NVIDIA's GB200 NVL72 rack-scale systems, GPT-5.5 delivers a 35x reduction in cost per million tokens and a 50x improvement in token output per second per megawatt. Its 1-million-token context window enables complex, long-form understanding and multi-step agentic tasks. This release significantly disrupts the competitive landscape, putting pressure on rivals and expanding possibilities for developers and enterprises.