On Wednesday, Amazon introduced Nova Premier, touted as the most advanced AI model in the Nova series. This latest iteration can process text, images, and video, although it does not handle audio. Available through Amazon Bedrock, their AI model development platform, Nova Premier is designed to tackle complex tasks that require a nuanced understanding of context, multi-step planning, and accurate execution across various data sources.
The Nova family of models was first unveiled during AWS re:Invent last December, and since then, the range has expanded to include models capable of generating images and videos, as well as those that can understand audio. While Nova Premier boasts a substantial context length of 1 million tokens—allowing it to process approximately 750,000 words at once—its performance on certain benchmarks falls short compared to leading models from competitors like Google. For instance, on coding assessments like SWE-Bench Verified, Premier lags behind Google’s Gemini 2.5 Pro, and it does not perform as well on math and science knowledge tests such as GPQA Diamond and AIME 2025.
Despite these drawbacks, Nova Premier shines in areas like knowledge retrieval and visual comprehension, where it has received favorable results according to Amazon’s internal evaluations. Pricing for the model in Bedrock is set at $2.50 per million input tokens and $12.50 for every million tokens generated, aligning it closely with Google’s Gemini 2.5 Pro pricing structure.
It is essential to note that Nova Premier is categorized as a model that does not engage in reasoning. Unlike OpenAI’s models, which can take additional time for careful consideration and verification of outputs, Premier focuses more on efficiency than exhaustive accuracy.
Amazon is advocating for the use of Nova Premier in training smaller models through a process called distillation, where its capabilities can be transferred for specific applications, thereby creating more efficient models. The company views AI as a vital component of its growth strategy, with CEO Andy Jassy indicating that Amazon aims to develop over 1,000 generative AI applications. He also highlighted that Amazon’s AI revenue is experiencing significant growth, achieving triple-digit increases year-over-year and resulting in a multi-billion-dollar annual revenue run rate.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
