OpenAI is making strides in AI accessibility with the introduction of GPT-4o Mini, a compact and much more affordable version of its advanced language model. This new model aims to make powerful AI technology available to a broader range of developers who previously found the cost of using larger models like GPT-4 prohibitive.
Building applications with OpenAI’s full-sized models has been expensive, pushing many developers towards cheaper alternatives such as Google’s Gemini 1.5 Flash and Anthropic’s Claude 3 Haiku. Recognizing the need for more affordable options, OpenAI has launched GPT-4o Mini to provide a cost-effective solution.
What is GPT-4o Mini?
GPT-4o Mini offers capabilities that surpass GPT-3.5 while being much cheaper to use. It’s now available to ChatGPT users on Free, Plus, and Team plans, and can also be accessed by developers through OpenAI’s API. This version supports both text and vision processing, and there are plans to add support for video and audio inputs and outputs soon.
How does it perform?
Despite being designed for simpler tasks, GPT-4o Mini shows impressive performance. It scored 82% on the MMLU benchmark, which is significantly higher than GPT-3.5’s 70% and comparable to other lighter models like Claude 3 Haiku and Gemini 1.5 Flash.
OpenAI has demonstrated the potential of GPT-4o Mini through partnerships with companies like Ramp and Superhuman. Ramp used the model to develop a tool for scanning receipts and extracting expense data from images. Superhuman incorporated it into their email client to provide AI-generated suggestions for email responses.
In summary, GPT-4o Mini offers a powerful and affordable option for developers, making advanced AI technology more accessible and opening up new possibilities for innovation.
