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.
Together AI's Jev-style classifier. A LoRA fine-tune of Qwen3.5-4B that reads a state, a question and 2 to 24 options and returns one option letter. Served on Together's platform; recipe published.
Decides
choice
choice, classify, route
Architecture
cua-s1
tev
Fine-tuned from
—
qwen/qwen3.5-4b
License
mit
Unspecified — weights licence being finalised
Availability
Open weights
Open weights + hosted API
Hosted by
—
Together AI
Input price
—
$0.042/MTok
Decision accuracy
100.0%
88.0%
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 tev1?
cua-s1-forms is from Cua and tev1 from Together AI. cua-s1-forms has open weights you can download and run; tev1 has open weights and a hosted API. Both answer choice questions. Only tev1 answers classify and route. cua-s1-forms is the smaller model, at 706K parameters to 4.0B. cua-s1-forms is licensed mit; tev1, other.
Which is more accurate, cua-s1-forms or tev1?
They report on different suites — cua-s1-forms 100.0% on 196-decision evaluation on real forms, tev1 88.0% on Together development set (reused, not held out) — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, cua-s1-forms or tev1?
cua-s1-forms: Free (open weights). tev1: $0.042 / $0 per 1M, or free to self-host. Open weights cost nothing per call beyond your own hardware.
Can I run cua-s1-forms or tev1 locally?
Yes, both: systemone pull cua/cua-s1-forms and systemone pull together-ai/tev1 download the weights.
