AI Model Size No Longer Determines Enterprise Choice

The AI industry has long been driven by the assumption that the largest model is the best. However, this assumption is now breaking down as enterprises prioritize cost, control, and task-specific capabilities. The reason for this shift is simple: model bills can reach millions of dollars a month, making the cheapest model that meets the quality bar the preferred choice.

Model routing has emerged as a solution to this problem. By automating the selection of the most suitable model for each task, model routing sends requests to the cheapest model that meets the quality bar. This approach has been made possible by the rise of specialized, industry-specific models that fill the gap left by the shift away from large, general-purpose models.

Gartner expects 40% of enterprise applications to embed task-specific AI agents by the end of 2026, up from under 5% a year earlier. This shift is driven by the economics of AI model development, with per-token prices collapsing and the industry moving towards a 'good enough' principle. In other words, most tasks do not need a frontier system, and buyers have worked out that most tasks can be handled by a cheaper model that meets the quality bar.