A 149.6M ModernBERT-base cross-encoder with a scalar head that picks an agent's next action or tool from labels and descriptions you supply. Each option is scored against the state and a softmax ranks the set; no text is generated. Experimental preview, not calibrated.
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, rank, route
choice, classify, route
Architecture
modernjev
nativ
Fine-tuned from
answerdotai/modernbert-base
dicta-il/neodictabert
License
apache-2.0
cc-by-4.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
73.4%
96.2%
Calibration error
—
0.010
Valid action rate
—
—
Median latency
—
—
p95 latency
—
—
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 modernjev-decide and nativ-he-decision?
modernjev-decide is from Maziyar Panahi and nativ-he-decision from Yoav Pinto. Both have open weights you can download and run. Both answer choice and route questions. Only modernjev-decide answers rank. Only nativ-he-decision answers classify. modernjev-decide reads up to 4K tokens of state, against 1K tokens for nativ-he-decision. modernjev-decide is the smaller model, at 150M parameters to 364M. modernjev-decide is licensed apache-2.0; nativ-he-decision, cc-by-4.0.
Which is more accurate, modernjev-decide or nativ-he-decision?
They report on different suites — modernjev-decide 73.4% on AgentToolDecisions-180K held-out test, next action type (1,158 decisions, 3 choices; the maker's split), nativ-he-decision 96.2% on nativ-bench user-intent task, 4 options (MASSIVE Hebrew, 2,973 items; the maker's own suite) — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, modernjev-decide or nativ-he-decision?
modernjev-decide: Free (open weights). nativ-he-decision: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run modernjev-decide or nativ-he-decision locally?
Yes, both: systemone pull maziyar-panahi/modernjev-decide and systemone pull yoav-pinto/nativ-he-decision download the weights.
Evaluation suite
AgentToolDecisions-180K held-out test, next action type (1,158 decisions, 3 choices; the maker's split)
nativ-bench user-intent task, 4 options (MASSIVE Hebrew, 2,973 items; the maker's own suite)
