The first System One model. Reads a state, answers typed Choice, Score and Noul questions in one call with calibrated probabilities, and generates no text. Closed weights, served by TypeSafe AI.
Invergent's decision model for text and images. A fine-tune of Gemma-4-26B-A4B (about 4B parameters active per token) that answers Choice, Noul and Score questions, with an optional thinking mode for harder questions.
Decides
choice, score, noul, classify, route
choice, score, noul, classify, route
Architecture
jev
rune
Fine-tuned from
—
google/gemma-4-26b-a4b-it
License
proprietary
apache-2.0
Availability
Hosted API
Open weights + hosted API
Hosted by
TypeSafe AI
Invergent
Input price
$0.042/MTok
—
Decision accuracy
—
—
Calibration error
—
—
Valid action rate
—
—
Median latency
—
—
p95 latency
—
—
Evaluation suite
Figures are from each model’s manifest; accuracy and latency are what the publishers report, on their own suites and hardware. Add a third model.
Questions
What is the difference between jev and rune?
jev is from TypeSafe AI and rune from Surogate (Invergent). jev is only available as a hosted API; rune has open weights and a hosted API. Both answer choice, score, noul, classify and route questions. rune reads up to 262K tokens of state, against 32K tokens for jev. jev is licensed proprietary; rune, apache-2.0.
Which is more accurate, jev or rune?
Neither publishes an accuracy figure. Test both on your own labelled examples.
Which is cheaper, jev or rune?
jev: $0.042 / $0 per 1M. rune: Hosted, price not published, or free to self-host. Open weights cost nothing per call beyond your own hardware.
Can I run jev or rune locally?
rune yes — systemone pull surogate/rune downloads its weights. The other is only served as a hosted API.
