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.
Native MLX FP16 conversion of Laya for Apple silicon. The same ModernBERT-large encoder, decision transformer and heads, running in MLX with no PyTorch; the checkpoint Laya Studio fine-tunes from.
Decides
choice
choice, score, noul
Architecture
cua-s1
laya
Fine-tuned from
—
convai-innovations/laya
License
mit
apache-2.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
100.0%
—
Calibration error
—
—
Valid action rate
—
—
Median latency
—
—
p95 latency
—
—
Evaluation suite
196-decision evaluation on real forms
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 laya-mlx?
cua-s1-forms is from Cua and laya-mlx from aac6fef. Both have open weights you can download and run. Both answer choice questions. Only laya-mlx answers score and noul. cua-s1-forms is the smaller model, at 706K parameters to 421M. cua-s1-forms is licensed mit; laya-mlx, apache-2.0.
Which is more accurate, cua-s1-forms or laya-mlx?
Only cua-s1-forms publishes an accuracy figure (100.0% on 196-decision evaluation on real forms); laya-mlx does not, so there is no comparison to make without your own test.
Which is cheaper, cua-s1-forms or laya-mlx?
cua-s1-forms: Free (open weights). laya-mlx: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run cua-s1-forms or laya-mlx locally?
Yes, both: systemone pull cua/cua-s1-forms and systemone pull aac6fef/laya-mlx download the weights.
