Jev
Jev by TypeSafe AI is a decision model: it answers typed questions with a probability for each allowed answer instead of generating text. It scores 60.1 on the Jev Decision Index (rank 8). Hosted, it costs about $0.0084 per 1,000 decisions.
Pricing
| Provider | Input / 1M | Output / 1M |
|---|---|---|
| typesafe | $0.04 | $0.00 |
| openrouter | $0.04 | $0.00 |
| vercel | $0.04 | $0.00 |
Cost per 1,000 decisions assumes 200 input tokens per decision.
Specifications
- Provider
- TypeSafe AI
- Parameters
- —
- Released
- Sep 2026
- Open weights
- No
- Question types
- Choice (one of N options), Score (ordered rubric), Yes / no
- Mechanism
- Hosted API
- Backbone
- —
- Jev-compatible request
- Yes
- Runs on CPU
- —
- License
- Proprietary (hosted API)
- Version
- jev-1.13.0
Benchmarks
- Jev Decision IndexECE 0.074
- 60.1
- DecisionBenchECE 0.128
- 0.720
- jev-bench (Jevify)
- 0.733
Jev Decision Index: public 58.0, ECE 0.074, Brier 0.356. Lower ECE means its probabilities can be taken at face value.
Speed & latency
Compare Jev with…
FAQ
How much does Jev cost?
Hosted, Jev costs $0.04 per 1M input tokens, about $0.0084 per 1,000 decisions at 200 input tokens each.
What questions can Jev answer?
Jev answers choice (one of n options), score (ordered rubric), yes / no questions, returning a probability for each allowed answer.
Is Jev well calibrated?
On the Jev Decision Index its expected calibration error (ECE) is 0.074; below about 0.05, a stated 80% confidence is right roughly 80% of the time.
Is Jev open weight, and can it run on a CPU?
No, Jev is closed and served as an API.
Data as of 2026-10-14. Source: Metatext System One registry (metatext.io/system-one). Pricing is per 1M tokens (USD); confirm with the provider before production use.