“Image 4-To-Tage” text from Google is a great boring pictures “

Google has unveil Her latest model from the text to Imagen 4 image with the usual promise of “a great improved text” on the previous version, Imagen 3. The company also presented a new version of Deluxe called IMAGEN 4 Ultra designed to follow more accurate text claims if you are ready to pay more. Both reach a paid pace in Gemini API And to conduct limited free tests in Google Ai Studio.

Google describes the main Imagen 4 model as “going to most tasks” at $ 0.04 per image. Imagen 4 Ultra, at the same time, is “when you need your photos to follow the instructions accurately” with the promise of “strong” output results compared to other photo generators like Dall-E and Midjourney. This model enhances the price by 50 percent to $ 0.06 per image.

The company displayed a set of photos including a three -panels created by IMAGEN 4 ULTRA showing a small space ship being attacked by a giant blue lizard … with some sound effects such as “Crunch!” And the lack of concept, “it was !!” The image followed the rhythm listed for Beat and looked fine, unlike to present from the 3D application.

Another directed reading “In front of an ancient travel card for Kyoto: iconic Paguda under cherry flowers, snow -covered mountains at a distance, clear blue sky, vibrant colors. ” Imagen 4 goes out to “T”, although in a general style lacks any magic. Another picture showed a long -distance walking couple waving from the top of a rock and another, and a fictitious “Avant Garde” fashion. The pictures were definitely good quality and followed text claims accurately but still looked high.

Imagen 4 is fine and it seems that it has a moderate improvement before, but I am not completely impressive-especially market leaders, Dall-E3 and Midjourney 7. In addition, after the initial rush of enthusiasm, the audience appears to have been ill from AI ART, with the main use apparently ads sent on social communication or in the upper part of the materials.