whatmodel.sh / decision / Open-Jev-9B (JevBench mini)
Open-Jev-9B (JevBench mini)
↗ share
+ contribute
Open-Jev-9B (JevBench mini) is a source-available decision implementation from Zefan Cai.
Catalogue identity notes
JevBench 0.1.0 mini tested configuration open-jev-9b: Decoder, trained scorer head; base Qwen3.5-9B; readout rank-8 LoRA + scalar decision head; NLL + Brier. Separate identity because the report does not establish equivalence to the newer Decision Index checkpoint or serving recipe. No latency imported.
$ how it compares
The decision models ranked either side of Open-Jev-9B (JevBench mini), out of 357 with a score.
| # | model | score | where |
|---|---|---|---|
| 167 | JevK5 9B v0.3.3 | 54±1 | model files |
| 168 | Jevfire Qwen3.8-27B (decision readout) | 54±1 | source code |
| 169 | Open-Jev-9B (JevBench mini) | 54±2 | source code |
| 170 |  Solomon v1.1 | 54±1 | model files |
| 171 |  Sieve 4B | 54±1 | model files |
see all decision models · find the ones that fit your machine
$ what it's good at
consistency▃·▂▂·▅▃·▃▅▃█ #2 of 21 · 69% of its score · 5 tests
Trails JevAny-Gemma-4B (pointer) (JevBench mini) here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| JevBench mini 0.1.0 · Probability measure coherence (pass rate) | 50.2% ✎ | ▄·▁·▃▃▁·▃▄▁█ beats 16/20 | 53% |
| JevBench mini 0.1.0 · Representation consistency (pass rate) | 81.9% ✎ | ▃·▂·▂▂▂▃▅▅▃█ beats 19/20 | 13% |
| JevBench mini 0.1.0 · Logical coherence (pass rate) | 87.1% ✎ | ▃·▃·█▅▅██▃█▅ beats 20/20 | 2.3% |
| JevBench mini 0.1.0 · Choice-set coherence (pass rate) | 100.0% ✎ | ▁▃▄█▄▄·····▄ beats 20/20 | 1.2% |
| JevBench mini 0.1.0 · Batch independence (pass rate) | 100.0% ✎ | ▁··········█ beats 20/20 | 0.0% |
other tests1 not in a task · 28% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| JevBench mini 0.1.0 · unchanged-answer accuracy | 88.4% ✎ | ▁·▂·▂···▂▂▁█ beats 14/20 | 28% |
$ where to run it
Implementation and setup instructions ↗
Published implementation and setup documentation verified. This entry describes a decision readout or evaluated configuration, not necessarily separate trained weights. Follow the referenced version and model settings; hardware compatibility has not been verified. Source · checked 2026-10-31.
Model repository and documentation: github.com/Zefan-Cai ↗
$ how the score was worked out · JevBench 0.1.0 mini evaluatio… 97%
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 JevBench 0.1.0 mini evaluation reports | 6 | 97% |
| prior · pull toward the middle | 3% | |
| weighted average | = 54.0 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json