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.
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, score, noul
choice, score, noul, classify, route
Architecture
cygnet
decision
Fine-tuned from
google/gemma-4-12b-it
qwen/qwen3.5-9b
License
mit
apache-2.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
87.9%
77.4%
Calibration error
—
—
Valid action rate
—
—
Median latency
50 ms
—
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 cygnet and decision?
cygnet is from Blockbrain Labs and decision from vLLM Semantic Router. Both have open weights you can download and run. Both answer choice, score and noul questions. Only decision answers classify and route. decision is the smaller model, at 9.0B parameters to 12B. cygnet is licensed mit; decision, apache-2.0.
Which is more accurate, cygnet or decision?
They report on different suites — cygnet 87.9% on JevBench public set (231 items), JevBench CLI, 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, cygnet or decision?
cygnet: Free (open weights). decision: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run cygnet or decision locally?
Yes, both: systemone pull blockbrain-labs/cygnet and systemone pull vllm-semantic-router/decision download the weights.
