- Amazon plans to launch a new AI model with advanced reasoning capacities.
- The model aims to offer hybrid reasoning, a mixture of quick responses and a more complex reflection.
- Amazon favors external profitability and reference performance.
Amazon builds its own AI model which incorporates advanced “reasoning” capacities, learned Business Insider.
The offer should temporarily the launch by June under the Nova brand, a group of generative AI models, Amazon unveiled at the end of last year, according to a person directly involved in the project. This person asked not to be identified because he was not allowed to speak with the media.
Amazon wants the new model to adopt a “hybrid reasoning” approach which provides rapid responses and a more complex extensive reflection in a single system, added this person. An Amazon spokesperson did not respond to a request for comments.
Reasoning models have recently become the next AI border. They often work more slowly, but can also tackle more difficult problems by trying several solutions and turning back via chain techniques. Companies such as Google, Openai and Anthropic have recently published their own reasoning models, while Deepseek attracted a lot of attention to build a similar offer more effectively.
One of Amazon’s priorities is to make its Nova more economical reasoning model as a price than competitors, in particular the O1 of Openai, the Sonnet Claude 3.7 of Anthropic and the Flash Gemini 2.0 thought of Google, according to the person involved in the project.
Amazon previously declared that its existing internal models are at least 75% cheaper than third-party models available via its Rocky substratum development platform.
Another objective is to obtain it the next Amazon’s reasoning model classified in the top 5 for performance, based on external benchmarks that assess the development of software and mathematics skills, such as SWE, Berkeley’s function calling the ranking and likes it, among others, added this person.
This decision reflects Amazon’s commitment to invest in its own family of AI models, even if it preaches the need to offer a variety of choice of models via the foundation. Amazon’s AGE team, led by chief scientist Rohit Prasad, has worked on this new model.
He also puts Amazon in a more direct competition with Anthropic, the AI startup which has just launched its new model. Claude 3.7 Sonnet uses a similar hybrid approach, combining quick responses and longer chain output chain outings.
Amazon has invested $ 8 billion in Anthropic so far, and the two companies have been close partners, collaborating in areas like IA fleas and Cloud Computing.
Do you have a tip? Contact this journalist by email to ekim@businessinsider.com or signal, telegram or WhatsApp at 650-942-3061. Use a personal email address and a non-work device; Here is our guide to share information safely.
businessinsider
