<ul><li>OpenAI has revealed its next-generation reasoning model called o3, following the o1 family.</li><li>The o3 models are not yet available to the public but are available for testing by safety and security researchers.</li><li>The o3 family operates differently from traditional generative models and internally fact-checks responses for increased accuracy and reliability.</li><li>The o3 models have demonstrated significantly improved performance on benchmark tests compared to the o1 models.</li></ul>

OpenAI teases its ‘breakthrough’ next-generation o3 reasoning model

Discover more