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.
Juspay's open decision model. Qwen3.6-35B-A3B with a merged rank-16 LoRA, shipped with a pinned SGLang serving bundle that answers Choice, Score and Noul questions over text, images or a video through a TypeSafe-compatible /v1/systemone API.
Decides
choice
choice, score, noul, classify, route
Architecture
cua-s1
xor
Fine-tuned from
—
qwen/qwen3.6-35b-a3b
License
mit
apache-2.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
100.0%
90.0%
Calibration error
—
0.073
Valid action rate
—
—
Median latency
—
69 ms
p95 latency
—
162 ms
Evaluation suite
196-decision evaluation on real forms
JevBench public set (231 items), maker's self-run of Xor 1.2
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 xor?
cua-s1-forms is from Cua and xor from Juspay. Both have open weights you can download and run. Both answer choice questions. Only xor answers score, noul, classify and route. cua-s1-forms is the smaller model, at 706K parameters to 35B. cua-s1-forms is licensed mit; xor, apache-2.0.
Which is more accurate, cua-s1-forms or xor?
They report on different suites — cua-s1-forms 100.0% on 196-decision evaluation on real forms, xor 90.0% on JevBench public set (231 items), maker's self-run of Xor 1.2 — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, cua-s1-forms or xor?
cua-s1-forms: Free (open weights). xor: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run cua-s1-forms or xor locally?
Yes, both: systemone pull cua/cua-s1-forms and systemone pull juspay/xor download the weights.
