OpenAI GPT-4o mini: The maker of ChatGPT, OpenAI has confirmed that the company has launched a cost-effective tiny AI model GPT-4o mini, which will make this technology more reasonable and less energy-consuming, letting the startup focus on a wider pool of customers. OpenAI supported by Microsoft is the front-runner in the AI software market, which has been trying to make the technology available easily at a reasonable cost to developers to create apps that are powered by its model. The company has taken this decision at a time when affluent rivals such as Google and Meta are trying to seek a bigger share of the market.

COMMERCIAL BREAK
SCROLL TO CONTINUE READING

The officials of OpenAI have informed that GPT-4o mini is more than 60 per cent inexpensive compared to the GPT-3.5 Turbo which has been priced at 15 cents per million input tokens and 60 cents per million output tokens. As per the latest report, it presently outclasses the GPT-4 model on chat inclinations and has been able to score 82 per cent on Massive Multitask Language Understanding (MMLU). 

MMLU is a written intelligence and reasoning standard utilized to assess the competencies of language models. A greater MMLU score indicates that it can comprehend and utilize language in a better way across a wide range of domains, improving real-world usage. The AI technology maker says that the GPT-4o mini model's score matched 77.9 per cent for Google's Gemini Flash and 73.8 per cent for Anthropic's Claude Haiku.

Experts say that smaller language prototypes need fewer computational power to function, which makes them a more reasonable choice for firms with restricted resources that wish to implement generative AI in their processes. The officials of OpenAI say that ChatGPT's Free, Plus, and Team users will have the option to use GPT-4o mini as an alternative for GPT-3.5 Turbo, with enterprise users having access to it beginning next week.