← All stories
● Covered by 1 source · 1 reportMedium impact1 neutral

Decision Models Gain Traction; Microsoft Releases Its Own, Sqwish Raises $3M Pre-Seed

🔄 Updated 4d ago
New to BrevFeed? We gather this story from every outlet covering it into one summary — ranked by real-world impact, not just the latest headline — so you never miss what matters. What is BrevFeed? →

Key points

  • Decision models are LLMs that make quick choices from preset options.
  • Over 25% of Vercel AI router requests in October were for decision models.
  • Microsoft released its decision model, Microsoft-Decision-1.1, on Azure.
  • Sqwish raised $3 million pre-seed for its customizable decision model.

Rise of Decision Models

Decision models, a specialized form of large language models (LLMs) that do not produce output tokens, are gaining significant usage. These models are designed to quickly and cost-effectively choose between predefined options, indicating their confidence level. This functionality makes them suitable for production environments where rapid decision-making is prioritized over detailed reasoning.

In October, decision models constituted more than one in four requests on Vercel's AI model router, a notable share given their recent emergence. TypeSafe's Jev family of models has been a dominant player in this category, with OpenRouter handling over 1.5 billion queries to TypeSafe's models in a recent week.

New Entrants and Customization Focus

Following the introduction of Jev, numerous decision models have been developed, including offerings from major hyperscalers. Microsoft CEO Satya Nadella announced the immediate wide rollout of Microsoft-Decision-1.1 on Azure.

In parallel, Sqwish, a UK-based AI startup, secured a $3 million pre-seed funding round for its decision model, kettei fat. Sqwish's co-founders, Ushnish Sengupta and Federica Freddi, adapted their small LLM architectures to create their decision model.

Data and Post-Training Customization

Data is a critical component for effective decision models. TypeSafe CEO Diogo Almeida stated that calibrating Jev's judgment using statistically-adjusted synthetic data was a key challenge. Sqwish also utilizes synthetic data but differentiates itself by offering built-in post-training tools that allow customers to use their own production data to refine the model's judgment.

Sqwish CEO Ushnish Sengupta highlighted that this customization capability, enabling models to improve with customer-specific data, is a core aspect of their offering.

✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →

The daily brief

One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.

One email a day. Unsubscribe in one click, any time.

Today's brief

Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.

~4 min · 3 stories · Nov 21

▶ Play today's brief Listen on Spotify

New every morning, and the back catalogue is archived by date.

Reporting from

Decision models, a type of LLM optimized for quick choices between preset options, are seeing increased adoption, accounting for over 25% of requests on Vercel's AI model router in October. Microsoft has released its own decision model, Microsoft-Decision-1.1, on Azure, while AI startup Sqwish raised a $3 million pre-seed round for its customizable decision model.