OpenAI launches new super-efficient, low-latency ‘GPT-4o mini’

The new model purportedly offers comparable performance at a fraction of the cost.