Artificial intelligence is becoming the core driving force of the new generation of technological and industrial revolutions. And the large models have marked the beginning of a new path towards artificial general intelligence. Meanwhile the rapid development of artificial intelligence has brought significant changes and opportunities to scientific research. AI for science as a new paradigm, has more potential to accelerate the pace of innovation than ever before, and has become a frontier hotspot with remarkable achievements. In this article, it is elaborated in the first chapter that the development history of artificial intelligence and the technological evolution of large models, as well as the transformation of scientific paradigms driven by “big data + AI”. Then the state of the art in large models for science is systematically presented, including mathematics, biomedicine, chemistry, materials, and earth science. Furthermore, the challenges faced by artificial intelligence large models for science is analyzed. Finally, the recommended development paths of artificial intelligence large models for science are proposed.