OpenAI's latest multimodal language model, known as o3, has demonstrated an advanced ability to geolocate images with remarkable precision. Users report that o3 can analyze visual elements such as topography, vegetation, architecture, ridges, and contours to identify exact locations, even in complex environments like the Pyrenean valleys where hundreds of similar landscapes exist. This capability surpasses other multimodal large language models, which have struggled with such tasks. The model achieves this by zooming into images, cropping, enhancing details, and conducting rapid searches before providing precise map coordinates along with explanations of its reasoning. This development highlights a significant advancement in AI's visual reasoning and geographic identification skills.