OpenAI released a smaller and more affordable version of the technology that powers ChatGPT on Thursday. This development is part of the company's efforts to become more economically successful, known for its groundbreaking research in the field of artificial intelligence.

The new model, called GPT-4o mini, is 60% cheaper to operate than the previous model GPT-3.5 turbo, which until recently powered ChatGPT, according to OpenAI. At the same time, GPT-4o mini achieves higher scores in internal tests measuring intelligence and effectiveness.

OpenAI is the most well-known brand in the booming field of generative artificial intelligence. The fact that the company is now focusing on efficiency is a significant sign of a broader shift in the industry. Previously, OpenAI and competitors like Google and Microsoft primarily concentrated on developing the largest and most powerful AI models. Now, they are attempting to combine these efforts with the release of smaller models that are more profitable for them and more useful for business customers who need AI technology only for specific tasks.

Google as well as the startups Anthropic, Mistral, and Cohere have also released smaller models this year. Microsoft has highlighted a family of small models named Phi, which, according to the company, are only 1/100th the size of the model behind ChatGPT at the time.

Many generative AI startups and new market entrants like Apple are also focusing on smaller models that are cheaper to build, as they typically require smaller datasets for training. These models also need less computational power and can, in some cases, be run directly on hardware like a smartphone.

OpenAI explained that GPT-4o mini can currently interpret text and image inputs and will later be able to scan and output audio and video inputs. The technology was made available to users of the free and paid versions of ChatGPT starting Thursday. Next week, it will also be available to enterprise customers.