TechFlow news, on March 4, according to Business Insider, Amazon is developing an artificial intelligence model with advanced reasoning capabilities, planning to launch it under the Nova brand before June this year. The model adopts a "hybrid reasoning" approach, delivering both quick responses and more complex, extended thinking within a single system.
Insiders revealed that Amazon's primary goal is to make its Nova reasoning model more cost-effective than competing models such as OpenAI's o1, Anthropic's Claude 3.7 Sonnet, and Google's Gemini 2.0 Flash Thinking. Amazon previously stated its existing internal Nova models are at least 75% cheaper than third-party models offered via its Bedrock AI development platform.
Another objective is to rank the upcoming reasoning model among the top five in external benchmark tests, which primarily evaluate software development and mathematical skills, including SWE, Berkeley Function Calling Leaderboard, and AIME. The project is led by Amazon's AGI team, headed by chief scientist Rohit Prasad.
