GPT-5.2
OpenAI's smartest model with breakthrough benchmark performance - first model above 90% on ARC-AGI
GPT-5.2
OpenAI • December 2025
Training Data
Up to August 2025
GPT-5.2
December 2025
Parameters
~1.7 trillion
Training Method
Advanced Unified Architecture
Context Window
400,000 tokens
Knowledge Cutoff
August 2025
Key Features
Breakthrough Benchmarks • 100% AIME 2025 • ARC-AGI Leader • FrontierMath Excellence
Capabilities
Reasoning: Maximum (>90% ARC-AGI)
Math: Outstanding (100% AIME 2025)
FrontierMath: 40.3%
What's New in This Version
First model above 90% on ARC-AGI-1, perfect 100% on AIME 2025, 10% improvement on FrontierMath
OpenAI's smartest model with breakthrough benchmark performance - first model above 90% on ARC-AGI
What's New in This Version
First model above 90% on ARC-AGI-1, perfect 100% on AIME 2025, 10% improvement on FrontierMath
Technical Specifications
Key Features
Capabilities
Other OpenAI Models
Explore more models from OpenAI
GPT-5.6 Sol
OpenAI's flagship GPT-5.6 model — its most capable system yet for complex coding, cybersecurity, and scientific reasoning, with new max reasoning effort and ultra subagent mode
GPT-5.6 Terra
The balanced mid-tier of the GPT-5.6 family — matches GPT-5.5 quality on high-volume business tasks at roughly half the price
GPT-5.6 Luna
The fastest, most cost-efficient GPT-5.6 model — strong everyday capability at OpenAI's lowest price for summarization, drafting, and routine automation
