Turbo Instruct

GPT-3.5 Turbo Instruct: OpenAI's high-efficiency, task-oriented LLM, fine-tuned for precise instruction-following and direct text completion.

This model, officially `gpt-3.5-turbo-instruct`, is a dedicated instruction-following language model from OpenAI, distinct from its chat-optimized sibling, GPT-3.5 Turbo. It was engineered to replace older, less efficient instruct models—specifically the `text-davinci` series—while maintaining the speed and cost-effectiveness of the Turbo family. The model leverages Reinforcement Learning from Human Feedback (RLHF) for superior instruction adherence and reduced toxicity. It operates with a 4K token context window, excelling in direct Q&A, content generation, and code tasks where precise, non-conversational output is required.

https://platform.openai.com/docs/models