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.
Respan's behaviour-scoring model for evals, guardrails and monitoring. For each plain-language behaviour you define, it reads a conversation or agent trace and returns the probability the behaviour is present, absent or not observable, in one forward pass.
Decides
choice
noul, classify
Architecture
cua-s1
span
Fine-tuned from
—
—
License
mit
proprietary
Availability
Open weights
Hosted API
Hosted by
—
Respan
Input price
—
$0.020/MTok
Decision accuracy
100.0%
—
Calibration error
—
—
Valid action rate
—
—
Median latency
—
—
p95 latency
—
—
Evaluation suite
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 span-01?
cua-s1-forms is from Cua and span-01 from Respan. cua-s1-forms has open weights you can download and run; span-01 is only available as a hosted API. Only cua-s1-forms answers choice. Only span-01 answers noul and classify. cua-s1-forms is licensed mit; span-01, proprietary.
Which is more accurate, cua-s1-forms or span-01?
Only cua-s1-forms publishes an accuracy figure (100.0% on 196-decision evaluation on real forms); span-01 does not, so there is no comparison to make without your own test.
Which is cheaper, cua-s1-forms or span-01?
cua-s1-forms: Free (open weights). span-01: $0.02 / $0 per 1M. Open weights cost nothing per call beyond your own hardware.
Can I run cua-s1-forms or span-01 locally?
cua-s1-forms yes — systemone pull cua/cua-s1-forms downloads its weights. The other is only served as a hosted API.
