A 2.8 MB byte-level transformer for GUI form filling. For each form element it returns one probability per typed option (fill an entity, check, click or skip), using jevlike's option-attention head.
Small open typed-decision model with calibrated probabilities, first-class Nepali coverage.
Decides
choice
choice, score, noul, classify, route
Architecture
cua-s1
mira
Fine-tuned from
—
jhu-clsp/mmbert-small
License
mit
apache-2.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
100.0%
61.3%
Calibration error
—
0.083
Valid action rate
—
100.0%
Median latency
—
30 ms
p95 latency
—
43 ms
Evaluation suite
196-decision evaluation on real forms
s1-decision-bench
Latest version
2026.09
0.1.5
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 cua-s1-forms and mira?
cua-s1-forms is from Cua and mira from SAGEA. Both have open weights you can download and run. Both answer choice questions. Only mira answers score, noul, classify and route. cua-s1-forms is licensed mit; mira, apache-2.0.
Which is more accurate, cua-s1-forms or mira?
They report on different suites — cua-s1-forms 100.0% on 196-decision evaluation on real forms, mira 61.3% on s1-decision-bench — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, cua-s1-forms or mira?
cua-s1-forms: Free (open weights). mira: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run cua-s1-forms or mira locally?
Yes, both: systemone pull cua/cua-s1-forms and systemone pull sagea/mira download the weights.
