GPT-5.5 is a flagship large language model (LLM) released by OpenAI on April 23, 2026. Positioned as a successor to GPT-5.4, the model focuses on improved contextual understanding, agentic execution, and higher token efficiency. Unlike the "Thinking" variant designed for heavy multi-step logic, the standard version—often referred to as the non-reasoning variant—is optimized for lower latency and direct responses while maintaining the core capabilities of the GPT-5 series.
The model is characterized by its ability to act as an autonomous agent, capable of planning and executing multi-step workflows. It shows significant gains in agentic coding and computer use, allowing it to navigate complex software stacks, diagnose architectural failures, and maintain state across large systems. On benchmarks such as Terminal-Bench 2.0 and OSWorld-Verified, the model has demonstrated expert-level performance in navigating sandboxed environments and operating real computer interfaces independently.
Technically, GPT-5.5 features a 1-million-token context window for API users, enabling the processing of entire codebases or massive document corpora in a single call. Despite its increased parameter scale relative to previous generations, the model matches the per-token latency of GPT-5.4 through architectural optimizations developed in collaboration with NVIDIA. These optimizations allow the model to use significantly fewer tokens to complete equivalent tasks compared to its predecessors, reducing the overall computational cost for complex projects.