ChatGPT’s image-generation feature gets an upgrade

During Livestream on Tuesday, Openai CEO Sam Altman announced the first great increase Chatgpt’s Over a year, the image of the image generation.

Chatgpt can now use the company GPT-4O Model to create and change images and images locally. GPT-4O, a long time based on the operating chatbot platform, but so far the model can only create and edit the text – not pictures.

Altman said GPT-4O native Imic generation today in Chatgpt and Sora, open AI video generation product, video video generation product in AI for a 200-month-old Pro Plan of the company. Openai, feature, ChatGPT and free users say it is also spreading using the company’s API service soon.

The description is a GPT-4O, which has a little longer than the generation model, which has a GPT-4O, effectively From 3To make Openai describes more accurate and detailed images. GPT-4O, including images, including images, including images, including them or background facilities, or can change the details of “Ini.”

To strengthen the new image of image, explained Openai Wall Street Journal This was a GPT-4O training of ownership information from companies such as “Open Information”, as well as Shutterstock.

Many generative AI vendors see training information as a competitive advantage, so they keep it and keep any information near the chest. However, the details of the training are the potential source of allegations of both IP allegations, and another region is emerging for companies.

“We respect the rights of artists, and we have pictures that imitate any living artist, Brad Lightcap, Openai’s Chief Operations Director, said in a statement in the magazine.

Openai offers creators an opt-form that requires the removal of their work from the training information. The company also says that this respects the desire to implement websching boars, including photos and websites, including pictures, websites.

The improved image of Chatgept, Gemini 2.0 Flash, one of the company’s flagship models, is followed by an experienced native image exit heights for the Gemini 2.0 Flash. The strong feature went viral on social media – but not necessarily for the best reasons. Twins 2.0 Flash Image Component emerged Hugs severalAllows people to remove watermarks and create pictures that describe copyrighted characters.

This article was updated in PT in 12 to include an open statement to Wall Street magazine around GPT-4O training information.