Core Insights - Amazon announced the expansion of its Nova product line at the 2025 re:Invent global conference, introducing four new Nova 2 models and multiple services, achieving industry-leading price-performance ratios in inference, multimodal processing, conversational AI, code generation, and agent tasks [1][3]. Nova 2 Lite - Nova 2 Lite is a fast, cost-effective inference model designed for everyday workloads, capable of processing text, images, and video inputs to generate text outputs. It outperformed or matched competitors in various benchmark tests: 13 out of 15 tests against Claude Haiku 4.5, 11 out of 17 tests against GPT-5 Mini, and 14 out of 18 tests against Gemini Flash 2.5 [1][3]. - Key capabilities of Nova 2 Lite include document processing, extracting key information from videos, code generation, providing fact-based answers, and automating multi-step agent workflows [1][3]. Nova 2 Pro - Nova 2 Pro is an intelligent inference model that handles text, images, video, and voice inputs, generating text outputs. It is suitable for complex tasks requiring high accuracy, such as agent programming, long-term planning, and complex problem-solving. In benchmark tests, it matched or outperformed competitors: 10 out of 16 tests against Claude Sonnet 4.5, 8 out of 16 tests against GPT-5.1, and 8 out of 18 tests against Gemini 3 Pro Preview [2][4]. - Both Nova 2 Lite and Nova 2 Pro feature built-in web search and code execution capabilities, allowing them to access the latest internet information and run code, ensuring responses are based on current facts rather than solely on training data [2][4]. Market Adoption - Thousands of enterprises are currently using the Nova models to support various applications, including producing high-quality content, automating multi-step tasks, and accelerating the development of AI agents [2][4].