An open Jev-style decision model by Andy Zhang: a LoRA adapter on Gemma 4 26B-A4B (also on Gemma 4 12B) that reads a state and returns a temperature-calibrated probability for every option of a yes/no, choice or score question in one forward pass.
A 364M Hebrew decision encoder on NeoDictaBERT: give it a state, a question and 2 to 7 free-text options and it returns a probability for each option in one forward pass, about 70 ms on a laptop CPU. Hebrew only.
Decides
choice, score, noul
choice, classify, route
Architecture
ajev
nativ
Fine-tuned from
google/gemma-4-26b-a4b-it
dicta-il/neodictabert
License
apache-2.0
cc-by-4.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
—
96.2%
Calibration error
—
0.010
Valid action rate
—
—
Median latency
49 ms
—
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 ajev and nativ-he-decision?
ajev is from Andy Zhang and nativ-he-decision from Yoav Pinto. Both have open weights you can download and run. Both answer choice questions. Only ajev answers score and noul. Only nativ-he-decision answers classify and route. nativ-he-decision is the smaller model, at 364M parameters to 26B. ajev is licensed apache-2.0; nativ-he-decision, cc-by-4.0.
Which is more accurate, ajev or nativ-he-decision?
Only nativ-he-decision publishes an accuracy figure (96.2% on nativ-bench user-intent task, 4 options (MASSIVE Hebrew, 2,973 items; the maker's own suite)); ajev does not, so there is no comparison to make without your own test.
Which is cheaper, ajev or nativ-he-decision?
ajev: Free (open weights). nativ-he-decision: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run ajev or nativ-he-decision locally?
Yes, both: systemone pull andy-zhang/ajev and systemone pull yoav-pinto/nativ-he-decision download the weights.
Sequential requests, the maker's own measurement
nativ-bench user-intent task, 4 options (MASSIVE Hebrew, 2,973 items; the maker's own suite)
