system-one-bench

~/system-one-bench / compare/baseten-joey-1-vs-typesafe-jev-1-14

Head to head

Joey 1 vs Jev 1.14

Joey 1 scores 73.3 and Jev 1.14 79.3 on the System One Index. Joey 1 is more accurate on 2 of 16 text datasets, Jev 1.14 on 14. Same cases, same request, same scorer.

Overall

Joey 1Jev 1.14
System One Index73.379.3
English72.978.8
Arabic73.580.3
Vision Index–85.1
Calibration error2.1%6.0%
Median latency–136 ms
Cost per 1,000 calls$0.004$0.028

Text datasets, accuracy

Joey 1Jev 1.14
AG News English85.089.0
SST-2 English86.095.0
SST-5 English40.560.0
DAIR Emotion English52.557.5
Banking77 English93.078.5
BoolQ English80.090.0
XNLI English English84.086.0
Typed decisions English62.474.2
XNLI French French74.776.7
XNLI Arabic Arabic64.771.3
MASSIVE scenario Arabic71.565.0
MASSIVE intent Arabic84.585.0
OSACT4 offensive Arabic83.087.0
Tweet sentiment Arabic66.076.0
MIRACL rerank Arabic62.087.5
MIRACL pairwise Arabic83.090.0

Image datasets, accuracy

Joey 1Jev 1.14
SST-2, as an image–93.5
Arabic sentiment, as an image–63.5
Oxford-IIIT Pet–88.5
VQAv2 yes/no–86.5
CountBench–93.5

Model pages: Joey 1 · Jev 1.14. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.