GPTGenerative Pre-trained Transformer — OpenAI's family of large language models from GPT-1 through GPT-5.-3.5 Turbo cut APIApplication programming interface — a defined way for programs to talk to each other or to a service. costs dramatically while improving instruction-following over base GPTGenerative Pre-trained Transformer — OpenAI's family of large language models from GPT-1 through GPT-5.-3. It became the default model for early copilots, chatbots, and RAG pipelines before GPTGenerative Pre-trained Transformer — OpenAI's family of large language models from GPT-1 through GPT-5.-4 pricing dropped — democratizing LLMLarge language model — a neural network trained on vast text to generate and reason about language. features in SaaSSoftware as a Service — applications delivered over the internet on a subscription basis. products.

GPTGenerative Pre-trained Transformer — OpenAI's family of large language models from GPT-1 through GPT-5.-3.5 Turbo made LLMLarge language model — a neural network trained on vast text to generate and reason about language. APIs affordable enough for mainstream SaaSSoftware as a Service — applications delivered over the internet on a subscription basis. features.

Cost per token dropped enough that startups could ship AI features profitably.