OpenAI introduced a smaller, more cost-efficient version of its most powerful model GPT-4o — and it could mean even more AI-powered apps.
Donald Trump ‘lovefest’ tees up Jamie Dimon for Treasury Secretary job consideration
The new model, GPT-4o mini, is smarter and more than 60% cheaper than GPT-3.5 Turbo, according to OpenAI, meaning it is more affordable for developers building applications on OpenAI’s models. GPT-4o mini is priced at 15 cents per 1 million input tokens — or the unit of information an AI model uses, for example, a word or phrase — and 60 cents per 1 million output tokens. That is equivalent to 2,500 pages in a book, OpenAI said.
“We expect GPT-4o mini will significantly expand the range of applications built with AI by making intelligence much more affordable,” OpenAI said in a statement.
On the Massive Multitask Language Understanding (MMLU) benchmark, which measures language models’ capabilities across 57 subjects including math and law, GPT-4o mini scored 82%, outperforming other small models and GPT-3.5 Turbo, according to OpenAI. The small model will replace GPT-3.5 Turbo in ChatGPT, and Free, Plus, and Team users have access to the new model now.
At its release, GPT-4o mini can support text and vision in the API, or application programming interface, which allows developers to build their own AI models off of OpenAI’s technology. In the future, GPT-4o will have support for text, image, video, and audio, OpenAI said. The small model has knowledge up to October 2023.
“We envision a future where models become seamlessly integrated in every app and on every website,” OpenAI said. “GPT-4o mini is paving the way for developers to build and scale powerful AI applications more efficiently and affordably.”
OpenAI unveiled ChatGPT-4o, in May, which is multimodal, meaning it can see, hear, and have real-time conversations.
>>> Read full article>>>
Copyright for syndicated content belongs to the linked Source : Quartz – https://qz.com/openai-small-version-gpt4o-mini-chatgpt-cost-efficient-1851598869
