ChatGPT now offers a feature that businesses have been requesting for a long time.
OpenAI announced on Tuesday that paying developers can now fine-tune ChatGPT’s latest GPT-4o model with custom data. So businesses can upload custom data sets to train the AI chatbot and tailor it to their needs, whether that’s coding or writing.
Related: How to Start a Multi-Million Dollar Company, According to an IBM Engineer Turned Founder
The new capability means that a regenerative fashion brand, for example, can fine-tune ChatGPT to answer questions about the brand’s unique process for sourcing fabric. Or, a fun ice cream company can feed ChatGPT information about its recipes.
As a whole, successful niche businesses can customize ChatGPT to reflect their individual needs.
OpenAI has already opened up fine-tuning for its other, less advanced, AI models, like GPT-4o mini.
Businesses upload their custom data to OpenAI and give the AI model about an hour or two to process it, OpenAI software engineer John Allard told Bloomberg. Text is only supported right now, not images or other media.
OpenAI has specific instructions available for businesses that want to fine-tune ChatGPT.
