IT之家 4 月 7 日消息，據外媒 Bleeping Computer 今日報導，OpenAI 正在測試一項新功能：為 GPT-4o 圖像生成模型Add a watermark。

ChatGPT has been making headlines lately, in large part because of this image generation model. Not only can it accurately generate images with text, but it can also create themVisually realistic work, including the art style of imitating well-known Japanese animation companies such as "Studio Ghibli".

Initially available only to ChatGPT Plus users, the ImageGen model is now fully availableAll usersOpening. AI researcher Tibor Blaho has found that OpenAI appears to be testing the "ImageGen" watermark for images generated by free users.

Sources also said that OpenAI has recently started experimenting with adding watermarks to images generated by free accounts.And as long as you subscribe to ChatGPT Plus, you can save images without watermarks。

According to IT House, OpenAI has said in a previous blog that the training data of this model covers a large number of images and texts, so it not only has good aesthetic effects, but also hasWide range of practicality。

OpenAI noted: "Through the joint training of a large number of online images and text, we have not only learned the relationship between images and language, but also mastered the logical associations between images. Combined with strong post-training, this model shows:Stunning visual expressionThe resulting images are not only practical and coherent, but also have a good Contextual comprehension。 ”

OpenAI also confirmed that the ImageGen API is being developed and that developers will be able to do so in the futureUse this to build your own application products。