Sakhanda Wire
NVDA $152.86 -2.10% MSFT $505.56 -2.40% GOOGL $280.34 -1.19% META $618.69 -1.23% AMZN $256.63 -0.43%
← Back to the news

Decision models gain traction as Sqwish offers customer-specific tuning

Decision models are rapidly gaining adoption as an alternative to conventional large language models. In October, more than one-quarter of requests routed through Vercel’s AI model router went to decision models. Rather than generating output tokens, these systems select among predefined options and indicate their confidence, making them suited to applications where speed and cost matter more than extended reasoning.

TypeSafe’s Jev models remain the most-used publicly visible option, while dozens of competitors have appeared. Microsoft has also begun broadly deploying Microsoft-Decision-1.1 through Azure. Microsoft says the model was designed to balance cost, response time, intelligence and calibration, rather than maximize benchmark scores.

U.K. startup Sqwish Labs is pursuing a more customized approach with its kettei fat model. The company announced a $3 million pre-seed funding round and offers post-training tools that let customers use their own production data to refine model performance. CEO Ushnish Sengupta said continual tuning is intended to distinguish the product from interchangeable alternatives. Sqwish’s model has performed well on some benchmarks, although the company says testing in specific customer settings is more meaningful.

Sqwish co-founder Federica Freddi expects decision models to work alongside more capable frontier LLMs, with the two systems handling different kinds of tasks.

Written by Sakhanda Wire AI from reporting by TechCrunch. Facts, quotes and full context are in the original article below.

Originally published by TechCrunch on

Read the original on TechCrunch ↗

The reporting and the image belong to TechCrunch and are credited and linked above; the text on this page is Sakhanda Wire's own summary.

← Back to the news

More stories

All the latest news