GPT-5.2 (xhigh)

GPT-5.2 (xhigh) is a flagship large language model developed by OpenAI, released in December 2025 as part of the GPT-5 model series. It represents the highest reasoning tier within the GPT-5.2 family, designed for complex analytical tasks, advanced problem-solving, and agentic workflows. The "xhigh" designation refers to the maximum reasoning effort setting available via the OpenAI API, which allocates increased compute to internal processing before generating a final response.

The model features a 400,000-token context window and natively supports multimodal inputs, including text and vision processing. Compared to lower reasoning tiers, GPT-5.2 (xhigh) is characterized by higher verbosity and improved instruction-following, particularly in specialized domains such as software engineering, scientific research, and large-scale data analysis.

Capabilities and Architecture

Built on a proprietary architecture, GPT-5.2 (xhigh) introduces refined context management techniques, including a feature described as "compaction" to maintain coherence across long-form inputs. Benchmarks indicate high performance in reasoning-heavy evaluations, such as the GPQA (Graduate-Level Google-Proof Q&A) index. While offering high intelligence, the model typically operates with higher latency and greater token usage than the "Instant" or "Standard" configurations in the same family.