CITATION — REFERENCE ENTRY
System One models — TypeSafe AI documentation
- Key
- typesafe2026systemone
- Authors
- TypeSafe AI
- Issued
- 2026
- Type
- webpage
- Container
- TypeSafe AI documentation
- Publisher
- TypeSafe AI
Raw CSL JSON
{
"URL": "https://docs.typesafe.ai/concepts/system-one",
"type": "webpage",
"title": "System One models",
"author": [
{
"literal": "TypeSafe AI"
}
],
"issued": {
"date-parts": [
[
2026
]
]
},
"accessed": {
"date-parts": [
[
2026,
11,
5
]
]
},
"language": "en",
"publisher": "TypeSafe AI",
"container-title": "TypeSafe AI documentation"
}
Claims
-
TypeSafe says its probabilities are optimised against outcomes to reflect uncertainty, but that calibration holds across groups of predictions and does not guarantee any single answer is correct.
"Calibration is measured across groups of predictions; it does not guarantee that an individual answer is correct."
-
System One models do not write replies, produce code or explain their reasoning.
"do not write replies, produce code, or generate explanations of their reasoning"
-
A System One model evaluates a state and returns typed answers with probabilities. TypeSafe's API has three question types: Choice (pick one option from a defined set), Score (place the input on a numeric scale) and Noul (a yes/no question answered with a probability). Several independent questions can be asked together about one state.
"a class of AI models built to make fast, structured decisions that software can use directly"
Available in