OpenAI introduces a smaller, more efficient GPT-4o mini model for enhanced ChatGPT interactions

OpenAI has unveiled a more compact and cost-effective version of its popular large language model, ChatGPT. The new model, GPT-4o mini, will allow developers to create AI-powered applications and services at a 60 percent lower cost compared to the previous smallest model, GPT-3.5 Turbo. However, the most significant impact will be felt by consumers, as GPT-4o mini will now be the standard for free users of ChatGPT, leading to a noticeable enhancement in their overall experience.

OpenAI claimed that the GPT-4o mini scored 82 percent on an industry benchmark called MMLU, which stands for Measuring Massive Multitask Language Understanding, and includes 16,000 multiple-choice questions on 57 academic topics. That’s slightly lower than GPT-4o, which scored 88.7 percent, but higher than GPT-3.5 Turbo, which scored only 70 percent. AI experts have cautioned against relying on such benchmarks to measure how intelligent AI systems are, but so far they are the only way to measure the performance of large language models.

The smaller versions of the flagship models give developers more flexibility in building AI-based applications. After all, not everyone needs or can afford to use the biggest and most powerful models for every use case. In May, Google announced Gemini 1.5 Flash, its own lightweight model that the company said was optimized for speed and efficiency. Other AI companies, such as Anthropic, also have smaller versions of their full-scale models.

The GPT-4o mini can currently receive and generate text and images, but the model will eventually be able to handle other types of content such as audio and video. And while GPT-3.5 Turbo will be retired from ChatGPT, developers will still be able to use it through OpenAI’s API to build apps and services with it—until it eventually retires as well.