OpenAI has announced a new feature yesterday that will allow their corporate clients to fine-tune their most powerful model GPT-4o. This means they can customise the AI model with their own company data. 

Despite the widespread AI adoption, revenue generation has proven to be harder for companies. 

Customising a model means that it can specialise on a specific subject area and answer questions around that thoroughly. 

Until now, OpenAI did not offer fine-tuning for any of its flagship models including the GPT-4. However, it has allowed users to fine-tune its smaller variants like the GPT-4o mini. 

(For top technology news of the day, subscribe to our tech newsletter Today’s Cache)

Olivier Godement, head of product at OpenAI’s API, hopes that this will push customers to make changes to its model by working with the company directly instead of using an external party. 

There are several tech companies which gives users the option to customise all their AI models. 

Customers have to upload all their data to OpenAI servers and then train it on their own data which usually takes an hour or two. The data must be textual and no other types of content.