6 Groundbreaking AI Models Reshaping Enterprise Intelligence

eWEEK content and product recommendations are editorially independent. We may earn money when you click on our partners’ links. learn more.

At re:Invent 2024, Amazon AWS launched Nova, a cutting-edge family of artificial intelligence models designed to revolutionize the way enterprises leverage artificial intelligence. The announcement marks a major leap forward in Amazon’s artificial intelligence development, highlighting six generative AI models designed to simplify the generation of text, image and video content. Enterprise customers can access these artificial intelligence tools through Amazon Bedrock.

“Within Amazon, we have about 1,000 next-generation AI applications running, and we have a bird’s-eye view of the problems that application builders are still trying to solve,” said Amazon Senior Vice President of General Artificial Intelligence. Rohit Prasad said in a statement. “Our new Amazon Nova model is designed to help internal and external builders address these challenges and deliver compelling intelligence and content generation while achieving meaningful gains in latency, cost-effectiveness, customization, information foundation, and agency capabilities. progress.

Meet rising models

The Nova series includes five currently available models and a sixth model, Nova Premier, expected to be released in the first quarter of 2025. artificial intelligence model Designed to meet a variety of enterprise needs, from text to advanced multimedia content generation.

Text-only models include Micro, Lite and Pro:

• Xinwei: The Micro is the smallest product in the range but has the lowest latency, making it ideal for fast text-to-text tasks. Its contextual view of 128K tokens ensures efficiency for tasks such as text summarization, translation, content classification, interactive chatting and brainstorming, as well as simple mathematical reasoning and coding.
• Nova Lite: The model accepts images, videos, and text as input, but only produces text. It can handle inputs up to 300K in length and can analyze multiple images or videos up to 30 minutes long in a single request.
• Nova Professional Edition: The Pro is the most powerful of the three, offering superior accuracy and multitasking capabilities. It also has a 300K token capacity, making it suitable for complex workflows.

Creative content models include Canvas and Reel:

• New canvas: Canvas specializes in image creation, producing visual effects based on text or image prompts. It’s designed for industries like advertising and entertainment, with tools for color adjustment and background editing.
• Nova Scroll: Reel is a video generation tool that generates short videos based on text and image prompts. Users can manipulate camera movements using natural language commands, making it an ideal tool for marketers and content creators.
• Rising Star Prime Minister: State-of-the-art models built for complex reasoning and fine-tuning customized AI solutions. Amazon Nova Premier is still training. It is scheduled to launch in early 2025.

What makes Nova different?

A standout feature of Nova is its focus on cost efficiency. Amazon Nova Micro, Amazon Nova Lite, and Amazon Nova Pro are at least 75% cheaper than Amazon Nova Micro, Amazon Nova Lite, and Amazon Nova Pro best performing model In the respective intellectual lessons of the Amazonian bedrock, Amazon says. They are also good at customization, allowing businesses to Fine-tuning the model Use their proprietary data to improve accuracy and relevancy.

By 2025, contextual windows on some Nova models will expand to support more than 2 million tokens, equivalent to more than 1.5 million words, marking another leap forward. In addition, AWS plans to launch a speech-to-speech model in early 2025, which promises to achieve natural-sounding conversions by interpreting tone and rhythm. An ambitious arbitrary model is expected to emerge later this year, combining input and output of text, speech, images and video.

Learn more about Other top companies redefining artificial intelligence and advance this dynamic technology.