GPT

Generative Pre-trained Transformer. OpenAI's frontier family of LLMs. The current generation is GPT-5, with Thinking and Pro variants for deeper internal reasoning and maximum capability. Each generation has cut hallucination rates further versus the previous one.

How it works

GPT-5 is a multi-model deployment: a router decides whether your query needs the standard model, the Thinking variant (longer chains of internal reasoning), or the Pro variant (highest capability, longest context).

Example

A customer-facing agent uses the standard GPT-5 by default for cost efficiency, automatically escalates to the Thinking variant when the query requires multi-step reasoning, and falls back to Pro for the rare cases that need maximum capability.

Related terms

LLM Token

Need to actually use GPT?

We build production AI systems that put these concepts to work. 30 minutes, we map your use case.