In a significant move to democratize artificial intelligence, OpenAI has launched GPT-4o mini, a cost-efficient AI model aimed at making advanced technology more accessible. Priced at just 15 cents per million input tokens and 60 cents per million output tokens, GPT-4o mini is over 60 percent cheaper than its predecessor, GPT-3.5 Turbo.
Backed by Microsoft, OpenAI is striving to make AI development more affordable and efficient for developers worldwide. This initiative comes at a time when major competitors like Meta and Google are intensifying their efforts to capture a larger share of the AI market.
Despite its smaller size, GPT-4o mini outperforms the GPT-4 model in chat preferences and achieved an impressive 82 percent score on the Massive Multitask Language Understanding (MMLU) benchmark. To put this into perspective, it surpasses Google's Gemini Flash which scored 77.9 percent, and Anthropic's Claude Haiku with 73.8 percent.
MMLU is a benchmark used to assess the language models' ability to understand and reason across various domains, indicating that GPT-4o mini excels in real-world applications. The efficiency of smaller models like GPT-4o mini means they require less computational power, making them an attractive option for businesses with limited resources looking to integrate generative AI into their operations.
Currently, GPT-4o mini supports text and vision through its application programming interface (API), with plans to expand to video and audio inputs and outputs in the near future. Starting Thursday, ChatGPT's Free, Plus, and Team users will gain access to GPT-4o mini, replacing GPT-3.5 Turbo, while enterprise users will see the new model available next week.
OpenAI's introduction of GPT-4o mini underscores its commitment to making AI technology more affordable and widely available, fostering innovation and enabling a broader spectrum of users to leverage the power of artificial intelligence.
Reference(s):
OpenAI unveils cheaper small model GPT-4o mini, less energy-intensive
cgtn.com
