Amazon announces a set of Nova AI models

Amazon’s cloud computing division, Amazon Web Services (AWS), announced a new family of foundation models called Nova to power generative AI applications at its re:Invent 2024 conference on Tuesday.

Amazon is rolling out four Nova models in total: Micro, Lite, Pro, and Premier. Each have their own capabilities and sizes. Amazon also introduced a new image generation model, called Nova Canvas; and a video-generation model, Nova Reel.

Foundation models serve as a strong base for machine learning tasks, and are often used for generative AI tasks, which include creating new content like images, text, and audio. Instead of training and developing artificial intelligence models from scratch, data scientists often use a foundation model as a starting point to develop their own machine learning models.

“Inside Amazon, we have about 1,000 GenAI applications in motion, and we’ve had a bird’s-eye view of what application builders are still grappling with,” Rohit Prasad, senior vice president of Amazon artificial general intelligence, said in a statement.

“Our new Amazon Nova models are intended to help with these challenges for internal and external builders, and provide compelling intelligence and content generation while also delivering meaningful progress on latency, cost-effectiveness, customization, information grounding, and agentic capabilities,” he added.

Micro is a text-only model that delivers the lowest latency of the group, meaning it can process text the fastest. Lite is its multimodal model that processes images, video, and text inputs. Pro has the best combination of accuracy, speed, and cost for range of tasks. Premier, meanwhile, is best for complex reasoning tasks and for use as a teacher for distilling custom models.

The Nova class is available in AWS Bedrock, which is its AI development platform.

“We’ve continued to work on our own frontier models, and those frontier models have made a tremendous amount of progress over the last four to five months,” Amazon CEO Andy Jassy said onstage. “We figured, if we were finding value out of them, you would probably find value out of them.”

Amazon also announced on Tuesday that it’s partnering with Anthropic—the AI startup Amazon has invested $8 billion in—to build an AI compute cluster that uses its Trainium 2 chips. “When completed, it is expected to be the world’s largest AI compute cluster reported to date available for Anthropic to build and deploy its future models on,” Amazon said in a press release.

https://www.fastcompany.com/91239742/amazon-announces-a-set-of-nova-ai-models?partner=rss&utm_source=rss&utm_medium=feed&utm_campaign=rss+fastcompany&utm_content=rss