Amazon is in the process of developing a new artificial intelligence (AI) model that emphasizes reasoning capabilities. This upcoming model is expected to be integrated into the company’s Nova AI family and will predominantly serve enterprise clients through Amazon Bedrock and Azure AI Foundry, rather than being aimed at general consumers. Once launched, this large language model (LLM) is set to compete with other reasoning models in the industry, including OpenAI’s o3-mini, Gemini 2.0 Flash Thinking from Google, and DeepSeek-R1.
Amazon Said to Be Building a Reasoning Model
A recent report by Business Insider indicates that Amazon is developing this reasoning model entirely in-house. Citing sources familiar with the project, the report mentions that Amazon is aiming to incorporate “hybrid reasoning” capabilities into the model.
Last month, Anthropic launched the Claude 3.7 Sonnet AI model, which similarly features hybrid reasoning. This concept refers to a large language model that can quickly provide standard intelligence responses while also generating more intricate answers through a test time compute method, allowing it to tackle complex issues.
According to the report, Amazon is planning to unveil this unnamed AI model in June, with a strong focus on ensuring cost-effectiveness. This approach is essential for maintaining a competitive advantage in an increasingly crowded market of reasoning-focused models.
Despite prioritizing cost reduction, Amazon reportedly aims to uphold high performance standards for the model. Sources suggest that the company intends for the model to rank among the top five on third-party AI evaluation leaderboards.
This ambition likely pertains to the Chatbot Arena leaderboard, previously known as the Large Model Systems Organisation (LYSYS), which serves as a crowdsourced platform for developers and AI users to rank various models based on their experiences.
At present, the leaderboard is topped by Grok-3-preview, with GPT-4.5-preview and Gemini 2.0 Flash Thinking following in second and third places, respectively.
Interestingly, this news about Amazon’s reasoning model emerges shortly after the company introduced Alexa+, the advanced AI version of its voice assistant, featuring enhancements such as contextual understanding and improved control over smart home devices.
