Amazon to Launch ‘Reasoning’ AI Model to Rival OpenAI & Anthropic

Amazon is preparing to launch a new AI model with advanced reasoning capabilities under its Nova brand, which was unveiled late last year, by June, according to Business Insider. This model is designed to compete with offerings from OpenAI, Anthropic, and Google, emphasizing hybrid reasoning, cost efficiency, and strong performance on industry benchmarks.

What is Hybrid Reasoning?

Amazon’s upcoming AI model will use a hybrid reasoning approach, which blends:

• Quick, straightforward answers for simple queries.

• More complex, multi-step reasoning for advanced problem-solving, using chain-of-thought techniques.

This approach aligns with recent industry trends, as Google’s Gemini 2.0 Flash Thinking, OpenAI’s o1, and Anthropic’s Claude 3.7 Sonnet have all focused on enhancing AI’s ability to reason through complex tasks.

Amazon’s Competitive Edge: Cost & Performance

Amazon is prioritizing cost efficiency, aiming to make its Nova reasoning model significantly cheaper than competitors. The company has previously claimed its in-house Nova models are at least 75% cheaper than third-party models offered through its Bedrock AI platform.

Additionally, Amazon wants its model to rank among the top 5 in external AI benchmarks for:

• Software development skills (SWE benchmark).

• Function calling accuracy (Berkeley Function Calling Leaderboard).

• Mathematical reasoning (AIME benchmark).

Amazon’s Expanding AI Ambitions

Despite positioning itself as a platform for multiple AI models through Bedrock, Amazon is heavily investing in its own AI development. The company’s AGI team, led by head scientist Rohit Prasad, is spearheading this reasoning model project.

Interestingly, Amazon’s new model puts it in direct competition with Anthropic, the AI startup in which Amazon has already invested $8 billion. The two companies have maintained a close partnership, working together on AI chip development and cloud computing advancements. Anthropic’s latest model, Claude 3.7 Sonnet, follows a similar hybrid reasoning strategy, combining quick responses with deeper, structured thinking.

What This Means

Amazon’s move signals a more aggressive push into proprietary AI development, even as it continues to promote third-party models on AWS Bedrock. With cost efficiency and strong benchmark performance as key selling points, Amazon is positioning its Nova reasoning model as a serious competitor to industry leaders like OpenAI and Google.

As reasoning models become the next major frontier in AI, Amazon’s investment in hybrid intelligence—balancing speed with complex thought processes—could reshape enterprise AI solutions. Whether Nova’s reasoning model can truly outperform its rivals remains to be seen, but Amazon’s entry into the reasoning AI race raises the stakes for the industry.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing, image, and idea-generation support from ChatGPT, an AI assistant. However, the final perspective and editorial choices are solely Alicia Shapiro’s. Special thanks to ChatGPT for assistance with research and editorial support in crafting this article.