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.
A 0.6B replica of the System One idea trained on game environments. A Qwen3-0.6B backbone with an attention-based Choice head that scores a dynamic candidate set for Maze, Snake, ViZDoom and position prediction, shipped with its full training pipeline.
Decides
choice, score, noul, classify
choice, noul, score
Architecture
hopper
nanojev
Fine-tuned from
qwen/qwen3.5-4b
qwen/qwen3-0.6b
License
Research and demo use only (training data includes RACE, non-commercial); serving code Apache-2.0
mit
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
68.5%
—
Calibration error
0.102
—
Valid action rate
—
—
Median latency
—
—
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 hopper and nanojev?
hopper is from HopitAI and nanojev from TianyuCodings. Both have open weights you can download and run. Both answer choice, score and noul questions. Only hopper answers classify. nanojev is the smaller model, at 600M parameters to 4.0B. hopper is licensed other; nanojev, mit.
Which is more accurate, hopper or nanojev?
Only hopper publishes an accuracy figure (68.5% on JevBench public hard tier (111 items, measured by the authors)); nanojev does not, so there is no comparison to make without your own test.
Which is cheaper, hopper or nanojev?
hopper: Free (open weights). nanojev: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run hopper or nanojev locally?
Yes, both: systemone pull hopit-ai/hopper and systemone pull tianyu-codings/nanojev download the weights.
—
Evaluation suite
JevBench public hard tier (111 items, measured by the authors)
