AWS has launched Nova Premier, its most advanced AI model to date, designed specifically for enterprise use cases. The new model, available via Amazon Bedrock, targets complex, multi-step workflows and supports model distillation, enabling smaller models to inherit its capabilities with improved efficiency and reduced cost.

Nova Premier boasts impressive capabilities, including multimodal support for text, image, and long-form video inputs, with a one-million-token context window equivalent to around 750,000 words. The model will support over 200 languages, making it a versatile tool for global enterprises. According to AWS, Nova Premier's applications will span financial analysis, software automation, and agentic tasks involving orchestration across tools and data layers.

Industry analysts praise Nova Premier's unique features, particularly its multimodal capabilities and model distillation support. Deepika Giri, head of research – BD & AI at IDC Asia/Pacific, notes that Nova Premier distinguishes itself by applying large language models (LLMs) to niche Agentic AI scenarios where performance and cost-efficiency are critical. "Its multimodal capabilities also expand its relevance across a wide range of enterprise use cases," she added.

While Nova Premier is touted as AWS's "most capable model," it trails key rivals in some third-party benchmarks. The model lags behind Google's Gemini 2.5 Pro in coding (SWE-Bench Verified) and scores lower on math and science evaluations such as GPQA Diamond and AIME 2025. However, Amazon's internal testing shows Nova Premier performs strongly in knowledge retrieval and visual reasoning, with scores of 86.3 on SimpleQA and 87.4 on MMMU.

Pricing for Nova Premier is in line with industry standards, with a cost of $2.50 per million input tokens and $12.50 for output tokens, comparable to Google's Gemini 2.5 Pro. This competitive pricing, combined with its advanced features, makes Nova Premier an attractive option for enterprises seeking to leverage AI in their operations.

A key feature of Nova Premier is its support for model distillation within Bedrock. This allows enterprises to generate synthetic data from Premier and fine-tune smaller models, such as Nova Pro, Lite, and Micro, for targeted applications. According to AWS, a distilled Nova Pro model increased API invocation accuracy by 20% while delivering similar output at lower cost and latency. The process eliminates the need for labeled training data and is positioned for edge deployments and use cases with resource constraints.

Industry experts see the launch of Nova Premier as a strategic shift for AWS. Amandeep Singh, principal analyst at the QKS Group, notes that Nova Premier signals AWS's move from a neutral model host to asserting foundational control in the GenAI value chain. "This isn't about building the biggest model, it's about owning the orchestration, pricing, and architecture," he said. By combining proprietary models with Bedrock's flexible interface, AWS strengthens its appeal as a sustainable AI stack for enterprises.

Currently, Nova Premier is available only to approved Bedrock users. As the AI landscape continues to evolve, it will be interesting to see how Nova Premier is adopted by enterprises and how it competes with other advanced AI models in the market.