Amazon to deliver a smart, cost-efficient reasoning model with Nova AI in June
But whether it can outperform industry leaders remains to be seen.
Artificial Intelligence is evolving fast, and the latest battleground is reasoning—getting AI to think more like humans by breaking down problems step by step. Amazon is now making a big move into this space with its upcoming AI model, set to launch under the Nova brand by June 2025.
According to reports, this model will feature “hybrid reasoning,” a technique that allows AI to provide both quick answers and more complex, multi-step solutions within the same system.
Reasoning models have gained attention because they significantly improve AI’s ability to handle complex tasks, especially in math, coding, and science. Unlike standard AI chatbots that generate responses based on pattern recognition, reasoning models take a structured approach, testing multiple possibilities before settling on the best answer.

You can think of it this way, regular AI chatbots pull answers from a 'well of knowledge' they've been trained on. The problem is, if the answer isn’t in that well, they struggle. Reasoning models, on the other hand, actually ‘think’ through problems, breaking them down step-by-step and think of an answer—more like how a human would.
Competitors like OpenAI’s o3-mini, Google’s Gemini 2.0 Flash Thinking, and DeepSeek’s R1 have already entered this space. Anthropic, in particular, set the bar high with its Claude 3.7 Sonnet model, which also uses a hybrid reasoning approach.

One of Amazon’s biggest priorities is making its model cost-efficient. The company claims its in-house Nova models are at least 75% cheaper than third-party models available on its AWS Bedrock platform. This could be a game-changer, considering DeepSeek has built a reputation for offering AI models at incredibly low prices.
But cost isn’t the only focus—Amazon also wants its reasoning model to rank in the top five on AI leaderboards, which evaluate AI performance in areas like software development and mathematical problem-solving.
Last year, Amazon made $574.8 billion in revenue per a report from Techloy so it definitely has the deep pockets for a project like this. Mixed with the company's experience in cloud computing this new model could shake up the AI landscape.
Whether it can outperform industry leaders remains to be seen, but one thing is clear—the race to build smarter AI just got even more competitive.