A 144M-parameter System One model on the multilingual mmBERT-small encoder. Takes a state, a question and 2 to 20 options and returns one decision with probabilities for choice, score and yes/no questions, on a CPU.
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.
Decides
choice, score, noul, classify, route
choice, score, noul, classify, route
Architecture
julia
rizzo-flow
Fine-tuned from
jhu-clsp/mmbert-small
xhtoken/spark-x2.5-4b
License
apache-2.0
apache-2.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
73.2%
64.8%
Calibration error
—
0.112
Valid action rate
—
—
Median latency
—
195 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 julia-1 and rizzo-flow?
julia-1 is from Supersonic Labs and rizzo-flow from Rizzo AI Academy. Both have open weights you can download and run. Both answer choice, score, noul, classify and route questions. julia-1 is the smaller model, at 144M parameters to 4.0B.
Which is more accurate, julia-1 or rizzo-flow?
They report on different suites — julia-1 73.2% on typed-decisions test set (400 cases, 2,000 questions), rizzo-flow 64.8% on LocalLLaMA/typed-decisions test split (400 cases, 2,000 decisions), Q8_0 — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, julia-1 or rizzo-flow?
julia-1: Free (open weights). rizzo-flow: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run julia-1 or rizzo-flow locally?
Yes, both: systemone pull supersonic-labs/julia-1 and systemone pull rizzo-ai-academy/rizzo-flow download the weights.
Evaluation suite
typed-decisions test set (400 cases, 2,000 questions)
LocalLLaMA/typed-decisions test split (400 cases, 2,000 decisions), Q8_0
