What’s the story
OpenAI has unveiled a new, more affordable artificial intelligence (AI) model, GPT-4o mini.
The company states that this model is 60% cheaper than its least expensive existing model and offers superior performance.
OpenAI’s new model, GPT-4o mini, is a result of improved model architecture and refined training data and regimen.
This development is part of OpenAI’s mission to make AI “as broadly accessible as possible,” according to Olivier Godement, a product manager at OpenAI responsible for the new model.
GPT-4o mini outperforms other small models
OpenAI claims that its new model, GPT-4o mini, surpasses other “small” models on the market in several common benchmarks.
The company has made significant strides in the cloud AI market with its chatbot, ChatGPT, which debuted in late 2022.
OpenAI allows users to access the large language model that powers ChatGPT, called GPT-4o, for a fee.
It also offers a less powerful model, GPT-3.5 Turbo, at about a tenth of the cost of GPT-4o.
GPT-4o mini scores high on benchmark exam
The GPT-4o mini scored 82% on the Measuring Massive Multitask Language Understanding (MMLU), a benchmark test consisting of nearly 16,000 multiple-choice questions across 57 academic subjects.
In comparison, GPT-3.5 scored 70%, GPT-4o scored 88.7%, and Google’s Gemini Ultra claimed the highest-ever score of 90%.
These scores provide a comparative measure of the model’s performance against other AI models in the market.
OpenAI’s GPT-4o mini to support multimodal inputs
The new GPT-4o mini will be available for ChatGPT users on Free, Plus, and Team plans starting today, with Enterprise users gaining access next week.
The lightweight model will support text and vision in the API and will soon handle all multimodal inputs and outputs like video and audio.
Despite its capabilities, the model is intended for simple tasks.
This move is part of OpenAI’s ongoing effort to make AI more accessible to a wider audience.
