Amazon's AI Brain Gets an Upgrade: New 'Reasoning' Model Coming Soon

Amazon is cooking up something new in its AI kitchen. By June, it plans to serve a fresh model under its Nova brand – one that can both sprint and marathon through problems.

This isn't just another chatbot. The new AI aims to mix quick responses with deeper thinking, using "hybrid reasoning" to tackle complex problems. It's like having both a fast-food drive-through and a gourmet chef in the same kitchen.

Amazon's playing it smart with the price tag. They're aiming to undercut competitors like OpenAI and Anthropic, while still landing in the top 5 for performance on key benchmarks. Their existing Nova models already run 75% cheaper than rivals on their Bedrock platform.

The timing is interesting, coming right after Anthropic's launch of Claude 3.7 Sonnet. It's a bit awkward, considering Amazon has already poured $8 billion into Anthropic. It's like funding your friend's restaurant while secretly perfecting your own recipe.

The project comes from Amazon's AGI team, led by Rohit Prasad. While Amazon preaches model diversity through its Bedrock platform, this move shows they're not content just being the marketplace – they want to be a star vendor too.

Why this matters:

• Amazon's strategy is cleverly two-faced: they're both the AI shopping mall and a store owner, hedging their bets in the AI race
• The push for cost-efficiency suggests we're entering a new phase of AI competition where price matters as much as performance – the tech equivalent of a high-end restaurant offering early bird specials