GPT-3.5 Turbo

GPT-3.5 Turbo is the model OpenAI put behind ChatGPT when it launched to the public in November 2022, and it became a standalone API product on March 1, 2023. It’s a cheaper, faster variant of the GPT-3 line, tuned with instruction following and reinforcement learning from human feedback rather than raw few-shot completion. The original API version had a 4,096-token context window; a 16,000-token variant followed for customers who needed longer conversations or documents.

Pricing dropped repeatedly during its life, from an initial $0.002 per thousand tokens down to roughly half that within a few months, which made it the default cheap option for chatbot and automation projects through 2023 and into 2024. It was eventually overtaken by GPT-4 class models for tasks needing strong reasoning, but its low cost and speed kept it in wide use for simpler classification, summarization, and chat applications well after GPT-4 shipped.