A rank-16 LoRA on Qwen3.5-4B built for the JevBench setting. A document, a policy and a question go in; a calibrated distribution over the option letters comes out of one forward pass. Research and demo use only.
One of the earliest Jev-shaped scorers. A sequence-classification head on Qwen3.5-4B-Base scores each (state, question, option) triple and softmaxes per question. Non-commercial licence.
Decides
choice, score, noul, classify
choice, score, noul, classify
Architecture
hopper
scorer
Fine-tuned from
qwen/qwen3.5-4b
qwen/qwen3.5-4b-base
License
Research and demo use only (training data includes RACE, non-commercial); serving code Apache-2.0
cc-by-nc-4.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
68.5%
70.7%
Calibration error
0.102
0.044
Valid action rate
—
—
Median latency
—
112 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 hopper and system-one-qwen3-5-4b-scorer?
hopper is from HopitAI and system-one-qwen3-5-4b-scorer from pngwn. Both have open weights you can download and run. Both answer choice, score, noul and classify questions. hopper is licensed other; system-one-qwen3-5-4b-scorer, cc-by-nc-4.0.
Which is more accurate, hopper or system-one-qwen3-5-4b-scorer?
They report on different suites — hopper 68.5% on JevBench public hard tier (111 items, measured by the authors), system-one-qwen3-5-4b-scorer 70.7% on author-reported — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, hopper or system-one-qwen3-5-4b-scorer?
hopper: Free (open weights). system-one-qwen3-5-4b-scorer: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run hopper or system-one-qwen3-5-4b-scorer locally?
Yes, both: systemone pull hopit-ai/hopper and systemone pull pngwn/system-one-qwen3-5-4b-scorer download the weights.
JevBench public hard tier (111 items, measured by the authors)
