Recent reports suggest Amazon is actively developing a sophisticated AI model with advanced reasoning capabilities. This new product, codenamed Nova, is slated for a June launch and is part of a series of generative AI models released by Amazon late last year. Sources close to the project indicate the model employs a "hybrid reasoning" approach, enabling both rapid responses and more complex reasoning within the same system.

Reasoning models are considered a cutting-edge technology in the AI field. While typically slower than other models, they can solve more complex problems by backtracking through multiple solutions and employing chain-of-thought techniques. Major players like Google, OpenAI, and Anthropic have already launched their own reasoning models, intensifying market competition. Amazon aims for its upcoming Nova reasoning model to offer a price-performance advantage, undercutting competitors such as OpenAI's o1, Anthropic's Claude 3.7 Sonnet, and Google's Gemini 2.0 Flash Thinking.

Robot Artificial Intelligence 2025

Image Note: Image generated by AI, image licensing provider Midjourney

A key highlight of Amazon's AI strategy is that its internal Nova model is reportedly 75% cheaper than third-party models offered through its Bedrock AI platform. Furthermore, Amazon plans to achieve top-five performance across several external benchmark evaluations for its new reasoning model, demonstrating its ambition and capabilities in this area. These benchmarks include external standards for software development and mathematical skills, such as SWE, the Berkeley Function Call Leaderboard, and AIME.

Rohit Prasad, the chief scientist leading the project, stated that Amazon is significantly increasing its investment and technological development, while simultaneously encouraging diverse model choices through Bedrock. It's noteworthy that Amazon maintains a strong partnership with Anthropic, which recently launched the Claude 3.7 Sonnet model. The two companies collaborate extensively in AI chips and cloud computing, with Amazon investing a substantial $8 billion in Anthropic.

Overall, Amazon's development of a reasoning model underscores its unwavering commitment to AI technology, further intensifying its competition with major players in the field.