OpenAI has released a fine-tuning function for its latest artificial intelligence (AI) model, ‘GPT-4-o’. As customized models using open source models are gaining popularity, the intention is to monopolize the B2B market by affirming one of the best model.
Bloomberg reported on the twentieth (local time) that OpenAI is introducing a fine-tuning service to GPT-4o. Through fine-tuning, firms can retrain GPT-4o with their very own data for specific tasks or subject areas.
Since GPT-4o provides natural conversations much like those of humans, it is going to be highly useful in AICC (AI call centers) that require voice support.
For instance, a skateboard manufacturer could train a model on the specifics of wheel and board maintenance after which use it as a customer support chatbot.
Particularly, that is the primary time that OpenAI has released its flagship model for fine-tuning.
Previously, fine-tuning was supported in ‘GPT-3.5 Turbo’, which gained popularity amongst developers. Recently, the fine-tuning function was also supported in ‘GPT-4o Mini’, a less expensive and simpler version.
“We’re focused on helping customers fine-tune probably the most powerful models more easily,” said Olivier Godeman, head of product for OpenAI’s APIs.
Bloomberg reported that “OpenAI’s move comes as competition for business AI products intensifies and firms face pressure to prove that their AI investments are paying off.”
Meanwhile, Antropic recently partnered with LLM Ops specialist Kaylant to enter the marketplace for corporate custom models.
French open source specialist Mistral AI has also announced its entry into the US market, targeting the customized model market.
Reporter Park Chan cpark@aitimes.com
