NVIDIA has introduced a family of frontier-class multimodal large language models (MLLMs) called NVLM, designed to rival the performance of leading proprietary and open-source models like OpenAI's GPT-4 and Meta's Llama 3.1. NVLM combines the power of large language models with image interpretation capabilities, enabling it to handle complex tasks that go beyond what a purely text-based or image-based model could achieve. Key features include state-of-the-art performance on vision-language benchmarks, improved text-only performance after multimodal training, and three architectural options optimized for different tasks. NVLM's dynamic high-resolution image processing and diverse training data contribute to its superior performance in various applications such as healthcare, education, business, finance, and content creation.