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.
The flagship of the vLLM Semantic Router team's Decision 1.0 family. Qwen3.5-9B with a shared candidate head returns a probability for every supplied answer to choice, yes/no and score questions, over a 16,384-token input.
Decides
choice
choice, score, noul, classify, route
Architecture
cua-s1
decision
Fine-tuned from
—
qwen/qwen3.5-9b
License
mit
apache-2.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
100.0%
77.4%
Calibration error
—
—
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 cua-s1-forms and decision?
cua-s1-forms is from Cua and decision from vLLM Semantic Router. Both have open weights you can download and run. Both answer choice questions. Only decision answers score, noul, classify and route. cua-s1-forms is the smaller model, at 706K parameters to 9.0B. cua-s1-forms is licensed mit; decision, apache-2.0.
Which is more accurate, cua-s1-forms or decision?
They report on different suites — cua-s1-forms 100.0% on 196-decision evaluation on real forms, decision 77.4% on vLLM-SR decision benchmark (54 tasks, 3,766 decisions, weighted) — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, cua-s1-forms or decision?
cua-s1-forms: Free (open weights). decision: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run cua-s1-forms or decision locally?
Yes, both: systemone pull cua/cua-s1-forms and systemone pull vllm-semantic-router/decision download the weights.
