A local, Jev-compatible decision model from Rizzo AI Academy. XHToken's Spark-X2.5-4B with a merged typed-decisions LoRA, run on llama.cpp; yes/no, choice and score questions share one prefill of the state and are read from the answer-letter logits. No text is generated.
A Gemma 4 12B fine-tune for typed decisions that also chats and reads images. Ships as GGUF for llama.cpp, holds a 64K context on a 16 GB GPU, and serves /v1/systemone next to /v1/chat/completions.
Decides
choice, score, noul, classify, route
choice, score, noul, classify, route
Architecture
rizzo-flow
winnow
Fine-tuned from
xhtoken/spark-x2.5-4b
google/gemma-4-12b-it
License
apache-2.0
apache-2.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
64.8%
85.7%
Calibration error
0.112
—
Valid action rate
—
—
Median latency
195 ms
143 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 rizzo-flow and winnow?
rizzo-flow is from Rizzo AI Academy and winnow from EldanRing. Both have open weights you can download and run. Both answer choice, score, noul, classify and route questions. rizzo-flow is the smaller model, at 4.0B parameters to 12B.
Which is more accurate, rizzo-flow or winnow?
They report on different suites — rizzo-flow 64.8% on LocalLLaMA/typed-decisions test split (400 cases, 2,000 decisions), Q8_0, winnow 85.7% on JevBench public subset (231 items), Q8_0 — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, rizzo-flow or winnow?
rizzo-flow: Free (open weights). winnow: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Which is faster, rizzo-flow or winnow?
By their publishers’ figures, winnow answers in about 143 ms at the median and rizzo-flow in about 195 ms — measured on different hardware, so treat it as a rough guide.
Can I run rizzo-flow or winnow locally?
Yes, both: systemone pull rizzo-ai-academy/rizzo-flow and systemone pull eldanring/winnow download the weights.
LocalLLaMA/typed-decisions test split (400 cases, 2,000 decisions), Q8_0
