Babbage-002

Overview

Babbage-002 is a GPT base model from OpenAI that can understand and generate natural language or code. It is not trained with instruction following, so it continues the text it receives rather than answering instructions, and it is built for the legacy text-completion interface instead of chat. OpenAI lists it as deprecated.

Babbage-002 can generate up to 16,384 output tokens per request. It accepts text input and returns text output. Its knowledge cutoff is September 1, 2021.

Capabilities

Babbage-002 takes text as input and returns text. As a base model, it works well when the prompt is the beginning of the content you want, such as an opening paragraph or a partial code snippet, and the model continues from there.

Pricing

Babbage-002 is billed per token: $0.40 per 1M input tokens and $0.40 per 1M output tokens. Babbage-002 is available through RealRelay's OpenAI-compatible API.

Use cases

Use Babbage-002 for plain text completion: continuing a document from its opening lines, drafting boilerplate in a fixed style, or finishing a partial code snippet.

It suits existing completion-style workloads that need a low token price and an output limit of up to 16,384 tokens per request. It is not designed for chat or instruction-following tasks.