Microsoft has introduced a new artificial intelligence video generation tool in its Bing mobile application, leveraging OpenAI’s Sora text-to-video framework. The Bing Video Creator, unveiled on Monday, offers a complimentary method for users to create brief video clips without the typical ChatGPT subscription fee of $20 per month for Plus account holders.
Microsoft stated, “Bing Video Creator embodies our commitment to making AI video creation widely accessible. Creativity should be simple and reachable, aiding your quest for answers. We are thrilled to enable everyone to transform their thoughts into captivating videos through AI technology.”
The Video Creator is now being rolled out internationally, except in China and Russia, for the Bing Search applications on both Android and iPhone. Microsoft has also indicated that desktop and Copilot Search functionalities will be added shortly. Users can access the video generation feature from the menu in the bottom right corner of the Bing application or by providing a description in the search bar.
Users can queue up to three video creations simultaneously and will receive notifications when their videos are ready. Microsoft offers a “Standard” generation speed free of charge for all users, alongside a “Fast” option that processes clips in mere seconds. Users receive ten Fast generations at no cost, after which they will need to revert to the Standard speed or utilize 100 Microsoft Rewards points for additional Fast generations.
The generated videos are five seconds in length and formatted in a 9:16 vertical orientation, with support for 16:9 format expected to be introduced in the future. Videos will be kept on the Bing app for up to 90 days and can be downloaded or shared across other platforms.
The sample videos made available by Microsoft differ notably from the standard set by other AI models, such as Google’s Veo 3. For instance, the movements depicted in the otter chef and the sweeping hamster show awkward transitions, and the characters sometimes appear overly cartoonish. Nevertheless, based on previous experimentation with Sora following its launch in December 2024, these examples seem to better reflect the video’s capabilities than the refined clips OpenAI shared during the initial previews of Sora.
