Capable of very simple tasks, usually the fasted model in the gpt-3 series, and lowest cost. Can be fine-tuned. Has a context window of 2,049 tokens and training data up to October 2019.
gpt-3