What's the difference between OpenAI's o3-mini and o3-mini-high?

The difference between OpenAI's o3-mini and o3-mini-high primarily lies in their reasoning effort levels. While o3-mini processes queries at a medium reasoning effort, providing a balanced trade-off between speed and accuracy, o3-mini-high is designed for tasks that require 'higher intelligence,' resulting in more thorough reasoning at the cost of slower response times^[2].

In practice, users can expect o3-mini-high to outperform o3-mini in terms of the accuracy of answers, especially in more complex scenarios, but this comes with increased latency^[1]. O3-mini delivers faster responses while maintaining a high level of reliability for various STEM-related tasks, matching or exceeding the capabilities of the o1 model under certain conditions^[3].