Today, Amazon announced the new Nova series of state-of-the-art foundation models that deliver a leading price-to-performance ratio and are available exclusively on Amazon Bedrock.
For the past two years, Amazon has lacked exclusive state-of-the-art foundation models on its platform. The new Nova family of models solves this problem, and there are different models in the family that will be suitable for a variety of AI application scenarios.
- Amazon Nova Micro - A text-only model with the lowest latency responses with a context length of 128K tokens. It also supports fine-tuning and model distillation for better accuracy. It is generally available today.
- Amazon Nova Lite - A very low-cost multimodal model that can process image, video, and text inputs to generate text output. It supports inputs up to 300K tokens in length and can analyze images or up to 30 minutes of video in a single request. It also supports fine-tuning and model distillation for better accuracy. It is generally available today.
- Amazon Nova Pro - A highly capable multimodal model that can process up to 300K input tokens. Amazon claims that this new model achieves state-of-the-art performance on key benchmarks. It supports an input context of 300K tokens and can be used as a teacher model to distill custom variants of Amazon Nova Micro and Lite. It is generally available today.
- Amazon Nova Premier - Amazon's upcoming most capable multimodal model for complex reasoning tasks, which is still in training. It is expected to be generally available in early 2025.
- Amazon Nova Canvas - An image generation model that can create images from text or images provided in prompts. It also supports editing images using text inputs and controls for adjusting color scheme and layout. It is generally available today.
- Amazon Nova Reel - A new state-of-the-art video generation model that can create high-quality 6-second videos from text and images. Amazon will support 2-minute videos in 2025. Amazon Nova Reel will support the ability to control visual style and pacing, including camera motion control using text prompts. It is generally available today.
The above Amazon Nova models are now available in Amazon Bedrock in the US East (N. Virginia) AWS region. Amazon Nova Micro, Lite, and Pro are also available in the US West (Oregon) and US East (Ohio) regions via cross-region inference.
Amazon mentioned that the new Nova family of models understands and generates content in over 200 languages, with strong capabilities in English, German, Spanish, French, Italian, Japanese, Korean, Arabic, Simplified Chinese, Russian, Hindi, Portuguese, Dutch, Turkish, and Hebrew.
With the Nova series, Amazon is making a significant step in the field of Generative AI, offering powerful and cost-effective solutions for developers.
1 Comment - Add comment