Amazon’s AI Model Nova Reel Now Generates 2-Minute Videos

Amazon has upgraded its AI video model, Nova Reel, enabling it to generate videos up to two minutes long. First introduced in December 2024, Nova Reel marked Amazon’s entry into the growing generative video market, competing with offerings from OpenAI and Google. The updated version, Nova Reel 1.1, produces multi-shot videos with a consistent visual style across scenes. Users can input text prompts up to 4,000 characters long to generate videos made up of six-second segments totaling up to two minutes. The new update also introduces a feature called “Multishot Manual.” This mode allows users to include a reference image along with a shorter prompt of up to 512 characters, giving them more control over how each shot appears. With an image size of 1280x720 pixels, the model can generate videos comprising up to 20 different shots. Nova Reel is currently accessible through Amazon’s AWS services, including the Bedrock development suite. While users must request access to use the tool, Amazon has clarified that these requests are automatically approved. As with many generative AI models, concerns have been raised about how Nova Reel was trained. These systems rely on large datasets of existing videos to learn patterns, and it’s unclear whether Amazon used copyrighted materials without permission. The company has not disclosed what data was used to train Nova Reel or how creators might opt out of having their content included in the dataset. However, Amazon has stated it will protect AWS users from copyright claims related to media created with its AI tools under its existing indemnity policy. This move strengthens Amazon’s presence in the generative AI space, though transparency issues around training data remain. The new features, especially manual shot customization, give users greater creative control, positioning Nova Reel as a powerful tool for AI-generated video content.