Frontier Model
The most capable and computationally expensive AI models at any given time, typically requiring billions of dollars in training compute and representing the cutting edge of AI capabilities.
Frontier models are the most advanced large language models available, characterized by massive training costs (often tens to hundreds of millions of dollars per run), state-of-the-art capabilities, and deployment by major AI labs like OpenAI, Anthropic, and Google DeepMind. The term is used in policy discussions to distinguish these high-capability systems from smaller, more accessible models.
Also known as
frontier AI, frontier LLM, cutting-edge model, state-of-the-art model