Amazon Nova 2.0 Lite (low) is a configuration of the Nova 2 Lite multimodal foundation model optimized for low-effort reasoning. Part of Amazon's second-generation model family released in late 2025, it is designed to balance cognitive performance with high throughput and cost-efficiency. The "low" designation refers to the model's hybrid reasoning setting, which utilizes a reduced internal reasoning budget to deliver faster response times for everyday tasks that do not require exhaustive step-by-step thinking.
Built on a proprietary multimodal architecture, the model can process inputs across text, images, and video. It features a substantial 1 million token context window, allowing it to analyze extensive document sets, long-form video content, or complex codebases in a single prompt. While it is a "Lite" model in terms of resource consumption, it is benchmarked to perform competitively against other mid-sized frontier models in reasoning and information extraction tasks.
The model is integrated into the AWS ecosystem with features such as web grounding, allowing it to retrieve real-time information to improve response accuracy. It also supports native tool use and code execution, making it suitable for agentic workflows, document processing, and customer service automation. In its "low" reasoning mode, the model prioritizes efficiency, making it the most cost-effective tier within the Nova 2 Lite reasoning configurations.