Typed decisions from Google's frozen Gemma-4-12B-it, with no fine-tuning. A small shim over unmodified vLLM letters the options, reads the model's own probability for each letter at one answer position and applies one calibration temperature; one output token per decision.
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, score, noul
choice, classify, route
Architecture
cygnet
tev
Fine-tuned from
google/gemma-4-12b-it
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
87.9%
88.0%
Calibration error
—
—
Valid action rate
—
—
Median latency
50 ms
—
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 cygnet and tev1?
cygnet is from Blockbrain Labs and tev1 from Together AI. cygnet has open weights you can download and run; tev1 has open weights and a hosted API. Both answer choice questions. Only cygnet answers score and noul. Only tev1 answers classify and route. tev1 is the smaller model, at 4.0B parameters to 12B. cygnet is licensed mit; tev1, other.
Which is more accurate, cygnet or tev1?
They report on different suites — cygnet 87.9% on JevBench public set (231 items), JevBench CLI, 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, cygnet or tev1?
cygnet: 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 cygnet or tev1 locally?
Yes, both: systemone pull blockbrain-labs/cygnet and systemone pull together-ai/tev1 download the weights.
