What is GPT-4o mini? 🙏

SLMs more utility? Small and much cheaper.

Hey Everyone,

The race to more affordable SLMs in 2024 is off to a great start and especially this summer.

GPT-4o mini can currently take in and generate text and images, but the model will eventually be able to process other types of content like audio and video.

To get an idea how cheap GPT-4o mini is we can refer to this graph:

The new model, called GPT-4o mini, will cost developers 60 percent less to build AI-powered apps and services with as compared to GPT-3.5 Turbo, Open’s smallest model until about five days ago!

It scores an impressive 82% on the MMLU (Massive Multitask Language Understanding) benchmark, it surpasses the performance of its predecessor, GPT-3.5 Turbo, and even rivals some of the larger models in specific tasks

In fact a whole range of small language models arriving in mid 2024 are turning heads. Why? The new models—SmolLM, Mistral-Nemo, and GPT-4o Mini—each represent different approaches to creating more accessible AI, but they all share a common goal: bringing powerful language processing capabilities to a wider range of devices and applications.

Sure, GPT-4o mini can generate text, craft code and solve math problems much like its more capable namesake. However, the model does so slightly less accurately. But think of the accessibility improvements due to the cost.

NeMo is also fairly amazing by the way: