GLOSSARY

What is a System One model?

System One modelA System One model is a decision model: it reads a state and answers typed questions — choose an option, score on a scale, or give the probability that a condition holds — in one forward pass, with calibrated confidence and no generated text.

The name comes from Daniel Kahneman’s two systems of thought. System 1 is fast, automatic judgement; System 2 is slow, deliberate reasoning. A large language model that works through a problem in writing imitates System 2. A System One model does the other job: it returns the answer your code needs, directly.

The input is a state — text, a JSON document, a log, a conversation — and a set of typed questions about it. The output is one value per question, with a probability for every allowed answer. Nothing is generated, so there is nothing to parse and no way to answer off-schema.

TypeSafe AI released the first public System One model, Jev, in September 2026. Open models and reproductions followed within days, among them Laya, Kev, GLiNER2.5-Decide, CLM and Bespoke Nimble. The registry lists every one it knows about, with what each decides, its size, licence, price and reported benchmarks.

Example

{
  "answers": {
    "route": { "type": "choice", "choice": "technical", "probabilities": { "technical": 0.91, "billing": 0.06, "general": 0.03 } },
    "safe_to_run": { "type": "noul", "noul": 0.07 }
  }
}

Next term: System Two model · Every System One model

Example

Read more