GPT-3.5 Turbo Instruct

Overview

GPT-3.5 Turbo Instruct is an older OpenAI text model that OpenAI documents as compatible only with the legacy Completions endpoint, not with Chat Completions. It takes a text prompt and returns text, and it suits single-prompt completion work rather than multi-turn chat.

GPT-3.5 Turbo Instruct has a context window of 4,096 tokens and can generate up to 4,096 output tokens. It accepts text input and returns text output. Its knowledge cutoff is September 1, 2021, and it was released on September 28, 2023.

Capabilities

GPT-3.5 Turbo Instruct reads a text prompt and returns text, with a context window of 4,096 tokens shared by the prompt and the completion.

OpenAI documents it as working only with the legacy Completions endpoint, so it fits completion-style prompts rather than message-based chat.

Pricing

GPT-3.5 Turbo Instruct is billed per token: $1.50 per 1M input tokens and $2.00 per 1M output tokens. GPT-3.5 Turbo Instruct is available through RealRelay's OpenAI-compatible API.

Use cases

Use GPT-3.5 Turbo Instruct for single-prompt text completion, where one prompt goes in and one completion comes back, such as drafting a paragraph or extending a short passage.

It also suits existing integrations that were built on the legacy Completions endpoint and need to keep working without a move to chat-style messages.