OpenAI Makes GPT-3.5 Turbo Available for Fine-Tuning

Artificial intelligence company OpenAI is making its GPT-3.5 Turbo tool available to users for fine-tuning - an update developers and businesses have long asked for since the release of the model.

"This update gives developers the ability to customize models that perform better for their use cases and run these custom models at scale," OpenAI explained in a blog post. 

According to preliminary testing, a fine-tuned version of GPT-3.5 Turbo can even outperform GPT-4 when it comes to "certain narrow tasks." 

A private beta test also showed that fine-tuning the model improves its performance in terms of steerability, reliable output formatting, and customizing tones.

Additionally, fine-tuning GPT-3.5 Turbo allows it to handle 4,000 tokens at a time (double the number of tokens from previous fine-tuned models) and lets businesses shorten their prompts all while getting a similar performance.

"Early testers have reduced prompt size by up to 90% by fine-tuning instructions into the model itself, speeding up each API call and cutting costs," the company added.

GPT-3.5 Turbo launched earlier this year and is the most cost-effective option out of the GPT-3.5 family.

The model has been pre-trained on data up to September 2021 and has been optimized mostly for chat, but also works well for traditional completion tasks like API calls.

Edited by Nikola Djuric