Amazon Web Services (AWS), the cloud computing division of the tech giant, introduced the Nova artificial intelligence (AI) model on Tuesday at its ongoing RE:Invents Conference. There are five different large language models (LLMs) under the Nova brand, three of which can only provide text generation. In addition, Nova also includes image generation models and video generation models. The company says the new generation of AI models has improved intelligence and competitive prices and is currently available on Amazon bedrock.
AWS introduces Nova AI model
In one article, Amazon introduces the next generation of AI models in detail. Currently, five different LLMs have been introduced, as part of the Nova series, and the company’s CEO Andy Jassy stressed that the sixth AI model will be launched in 2025.
Of the five models, three – the new nova Micro, Nova Lite and Nova Pro can only produce text. However, there are differences between the three models. Micro accepts text only as input and provides the lowest latency response in the entire series. Its context window is 1,28,000 tokens.
Nova Lite, on the other hand, accepts images, videos, and text as input, but only generates text. The Nova Pro is the most capable multimodal AI model in the trio that can accomplish a wider range of tasks than the other two. The context window for both models is 300,000 tokens.
In addition, there are two other models in the Nova series, which Amazon calls the “creative content generation model.” First is Nova Canvas, an image generation model that accepts text and images as input. The company touted it as a tool for advertising, marketing and entertainment.
Finally, Nova Reel is a video generation model that generates short videos from text and image prompts. It also provides users with the option to use natural language prompts to control camera movement. All of these models are available for corporate customers of the company and can be used from Amazon bedrock platforms.
