Stealth Model Watch
Archive snapshot · Updated September 30, 2026
Jev — System One Model Archive
Jev is TypeSafe AI's first public System One model. Instead of generating text, it returns typed decisions with calibrated probabilities that software can act on directly.
Confirmed facts
Verified information
Sources marked as fact are publicly documented. Inference-tier items lack independent verification.
- Model ID
- Jev source ↗
- Developer
- TypeSafe AI, a San Francisco-based company founded in 2024 source ↗
- Announcement
- Introduced in the official post "Introducing System One Models & Jev" source ↗
- What it does
- Built for decisions inside software: it takes unstructured context and returns fast, structured decisions that applications can use directly source ↗
- Docs
- Official docs describe Jev as TypeSafe's flagship model and the first System One model source ↗
- Early access
- Reported as entering limited early access on September 15, 2026 source ↗
- Performance claims
- TypeSafe reports up to 200x faster inference and 400x lower cost than comparable LLMs — vendor-reported, independently unverified source ↗
- Licensing
- Described as a proprietary model source ↗
Why a non-stealth model is on this archive
Jev earns an archive here
TypeSafe launched Jev openly, but its 'System One' category is a new, watchable class.
Jev is not a stealth model — TypeSafe launched it openly. It earns an archive here because its "System One" category is a new, watchable class: models built for decisions rather than text. New model categories get tracked from day one on this site.
Direct answers
Jev FAQ
Short answers that preserve the same evidence boundaries as the facts above.
What is Jev?
Jev is the first public System One model from TypeSafe AI, built to return typed decisions with calibrated probabilities instead of generating text.
What is a System One model?
A model class built for decisions inside software — classification, routing, scoring, and safety checks — that skips token-by-token generation for speed.
How fast and cheap is Jev?
TypeSafe claims up to 200x faster inference and 400x lower cost than comparable LLMs. Independent benchmarks were not published at archive time, so treat these as vendor claims.
Is Jev open source?
It is described as proprietary. Open weights had not been announced at archive time.