Artificial intelligence

The Rise of Decision Models: Will Data-Driven Customizations Become the Next Trend?

Decision Models are experiencing rapid growth as a faster and less expensive alternative to Large Language Models, with a new trend emerging towards customizing these models using users' private data.

2026-11-17
3 min read
5 views
certi.news Editorial Team
The Rise of Decision Models: Will Data-Driven Customizations Become the Next Trend?

The AI environment is undergoing a remarkable shift, as Decision Models are capturing a significant share of intelligent computing tasks. According to data from the Vercel platform, requests for these models exceeded a quarter of total requests in October, reflecting an increasing reliance on tools designed to make quick choices among predefined options instead of generating long texts.

What are Decision Models?

These models differ from traditional Large Language Models (LLMs) in that they do not produce textual outputs; instead, they focus on using their general intelligence to make immediate, low-cost decisions, while providing a confidence level indicator for the decision made. These characteristics are vital in production environments that require ultra-fast response speeds that outweigh the accuracy of large, slow models.

The Competitive Landscape

The "Jev" model family from TypeSafe dominates this sector, having processed more than 1.5 billion queries in just one week. However, major tech companies (Hyperscalers) have entered the race strongly, with Microsoft announcing the launch of its own model Microsoft-Decision-1.1 via the Azure platform.

At the same time, startups are emerging as key players in this field; the British company Sqwish announced a $3 million funding round to develop its kettei fat model, focusing on tools that allow users to train models using their own data to improve judgment accuracy.

Why Does This News Matter?

Field experiments confirm that excessive reliance on performance benchmarks can be misleading in real-world work environments. Experts such as Achint Srivastava from Microsoft point out that the real challenge lies in balancing four dimensions: accuracy, reliability, speed, and cost efficiency.

What is Changing in Practice?

The focus is shifting from merely building general models to providing tools that allow organizations to customize models with their own data (Post-training tools). This trend gives companies a competitive advantage, as models become more skilled in specific tasks that matter to the user, reducing the risk of these technologies becoming mere digital commodities (Commodity).

Observers expect the future to witness an integration between "System 1 intelligence" (fast decision models) and "System 2 intelligence" (complex Large Language Models), where these systems work side-by-side to provide integrated technical solutions.

News source
TechCrunch AI
Open original source ↗
c
Author

certi.news Editorial Team

In the same category

You may also like

View all news