Decision Models Gain Rapid Adoption as New Frontier in AI Efficiency

The AI landscape is shifting toward 'decision models'—specialized LLMs designed to select between preset options quickly and cheaply without generating text tokens. This shift is evidenced by Vercel data showing decision models accounted for over 25% of requests on its AI router in October. While TypeSafe’s Jev family currently dominates the market, major hyperscalers are entering the fray, with Microsoft launching 'Microsoft-Decision-1.1' on Azure to compete on cost and latency. As the technology becomes increasingly commoditized, startups like UK-based Sqwish are focusing on customization through post-training tools that allow customers to refine models using their own production data. Industry experts suggest a future where these fast 'System One' decision models work in tandem with high-reasoning 'System Two' frontier LLMs, creating a multi-billion dollar market opportunity distinct from the massive valuations of general-purpose AI labs.

By Tim Fernholz · TechCrunch ·

Decision Models Gain Rapid Adoption as New Frontier in AI Efficiency

Tickers: $AMZN $MSFT

Topics: AI Infrastructure Decision Models Microsoft Azure Startup Funding TypeSafe Vercel

Read the full article at TechCrunch

More Technology news