One of the most pressing tasks of modern information technologies is to increase the speed of loading Internet pages. Google in solving this problem seeks to increase the efficiency of image compression, which carry the bulk of the transmitted information. Earlier, in 2010, Google already offered an original solution using the WebP format, which significantly, by 10 percent, reduced the size of the transmitted file while maintaining the quality. Given the fact that this format is used only by Google, it was decided to start optimizing images transmitted in JPEG format. The created algorithm was called Guetzli and allowed to reduce the size of JPEG files by 35 percent, and with the same file size allows you to transfer more detailed images. Google introduced the Guetzli algorithm (“cookies” in one of the dialects of Switzerland), the efficiency of which is ensured by the original Butteraugli model based on the psycho-visual characteristics of human vision. The model is based on the fact of a more effective visual perception by a person of the red and green color spectrum and the worst spatial vision of the blue spectral zone. Thus, compressing the image, you can neglect the blue color areas. The negative side of this method is the relatively low speed of image transformation. The original image on the left, in the center – libjpeg, on the right – GuetzliIn the case when users were given the opportunity to compare compressed images using traditional algorithms – libjpeg and the new method – Guetzli algorithm, the latter was most often preferred due to a sharper image and the absence of visual noise inherent to old techniques. All information about the new image compression technology is publicly available on a special GitHub service. At the same time, Google announced the full compatibility of compressed JPEG images with any browsers and other programs designed to view images.