Amazon has upgraded its generative AI video model, Nova Reel, to support the creation of videos up to two minutes long. The updated version, Nova Reel 1.1, was introduced on April 7, 2025, and now allows users to generate “multi-shot” videos with a consistent style across different shots, a feature that sets it apart in the competitive AI video space.
Originally launched in December 2024, Nova Reel made its debut as Amazon’s first entry into the growing generative video market, competing with similar models from OpenAI, Google, and others. The new version of Nova Reel 1.1 allows users to provide prompts with up to 4,000 characters that generate six-second shot videos from the model. This feature allows users to produce videos that contain a maximum of twenty distinct images as a result of the update.
The “Multishot Manual” mode, which was included in the most recent update, allows users to associate images with prompts that facilitate the creation of video shots. Through this mode, Nova Reel generates videos with a maximum resolution of 1280 by 720, which expands the potential for video customization.
The updated version of Nova Reel 1.1 is exclusively compatible with Amazon Web Services (AWS) platforms, such as Bedrock, which is Amazon’s AI development suite. Nova Reel necessitates that developers submit access requests via AWS; however, the platform automatically authorizes each request.
Despite its advancements, Nova Reel 1.1 raises concerns around the ethical implications of training generative models on large datasets, especially when videos are sourced without the consent of creators. While Amazon has not disclosed the sources of the training data, it has assured customers that it will protect them from any copyright infringement claims under its indemnification policy.
