OpenAI has introduced its latest generative artificial intelligence model, Sora, which can generate realistic videos up to a minute long from text prompts. The model has not been made available to the public yet, but OpenAI is granting access to red teamers and a limited number of visual artists, designers, and filmmakers to assess potential risks and gain feedback. OpenAI CEO Sam Altman showcased Sora’s capabilities by generating videos based on user prompts, including requests for a cooking session and a mythical creature flying through a sunset with a hamster. The announcement has generated excitement among users, who praised the progress made in less than a year.
OpenAI, the company behind ChatGPT and Dall-E, has developed Sora to generate realistic videos from text prompts. Although the model is not accessible to the public at the moment, it is being tested by red teamers and a select group of creative professionals for risk assessment and feedback. OpenAI CEO Sam Altman shared several videos generated by Sora, showcasing its potential. Users on the platform X were also invited to suggest prompts, leading to requests for unique scenarios like a cooking session hosted by a social media influencer and a mythical creature flying with a hamster in adventure gear.
The leading text-to-video AI model prior to Sora was developed by Runway, a Brooklyn-based company. Runway’s Gen-2 model, announced in March 2023, generated short and sometimes unsettling videos. Users were quick to notice the significant improvement in Sora’s videos and expressed their enthusiasm for the technology. OpenAI’s track record with ChatGPT and Dall-E, combined with the impressive capabilities of Sora, indicate the rapid progress made in AI development in less than a year. This announcement marks another milestone in OpenAI’s contributions to the field of artificial intelligence.
