OpenAI is launching a mini model of its newest AI mannequin.

These small AI fashions are supposed to be sooner and extra reasonably priced than the complete model — making them notably helpful for easy, high-volume duties. That ought to enchantment to smaller builders who don’t essentially have some huge cash for AI prices however wish to incorporate AI into their web site or app in a comparatively light-weight manner.

On this case, OpenAI introduced its newest flagship mannequin, GPT-4o, again in Might. The “o” stands for “omni,” a reference to the truth that the mannequin ought to have the ability to perceive speech and video, in addition to textual content.

And now there’s GPT-4o mini, which is the pared down model. GPT-4o mini at present helps textual content and pictures, however OpenAI says it would add video and audio capabilities sooner or later. This new mini mannequin is meant to be greater than 60% cheaper than GPT 3.5 Turbo, which it’s changing as OpenAI’s smallest mannequin. It additionally scores higher than competing small fashions on the MMLU (an business benchmark for reasoning).

Hit play to study extra, after which tell us what you assume within the feedback!