Amazon is reportedly working on an AI model that incorporates advanced “reasoning” capabilities, according to a recent report by Business Insider. This puts Amazon in the arena alongside competitors like OpenAI’s gpt-3.5-turbo and Chinese AI lab DeepSeek’s R1.
The report suggests that the model, potentially launching as early as June, could fall under the Nova brand. Amazon introduced Nova at its re:Invent developer conference last year. Reasoning models are designed to tackle queries in a more methodical, step-by-step fashion. This approach often leads to greater reliability, particularly in fields like mathematics and science.
Business Insider stated that Amazon plans to adopt a “hybrid” reasoning architecture for its new model, similar to Anthropic’s Claude 3.7 Sonnet, released a short time ago. This architecture may provide quick responses and more complex, in-depth analysis within a single, integrated system.
A key goal for Amazon is to make its Nova reasoning model more price-effective than its rivals, the report claims. However, achieving this could be a challenge, as DeepSeek has established itself with particularly competitive pricing.