system-one-bench

~/system-one-bench / compare/meragpt-decider-1-vs-typesafe-jev-1-13

Head to head

Decider 1 vs Jev 1.13

Decider 1 scores 61.8 and Jev 1.13 79.0 on the System One Index. Decider 1 is more accurate on 2 of 16 text datasets, Jev 1.13 on 14. Same cases, same request, same scorer.

Overall

Decider 1Jev 1.13
System One Index61.879.0
English67.478.5
Arabic54.779.9
Vision Index–83.9
Calibration error7.9%6.1%
Median latency–150 ms
Cost per 1,000 calls$0.012$0.028

Text datasets, accuracy

Decider 1Jev 1.13
AG News English87.588.5
SST-2 English93.094.5
SST-5 English55.562.0
DAIR Emotion English67.556.0
Banking77 English1.079.0
BoolQ English80.089.5
XNLI English English80.786.0
Typed decisions English74.272.7
XNLI French French67.376.7
XNLI Arabic Arabic67.369.3
MASSIVE scenario Arabic3.063.0
MASSIVE intent Arabic3.584.0
OSACT4 offensive Arabic86.090.0
Tweet sentiment Arabic64.076.0
MIRACL rerank Arabic81.086.5
MIRACL pairwise Arabic78.090.5

Image datasets, accuracy

Decider 1Jev 1.13
SST-2, as an image–94.5
Arabic sentiment, as an image–59.0
Oxford-IIIT Pet–83.5
VQAv2 yes/no–88.0
CountBench–94.5

Model pages: Decider 1 · Jev 1.13. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.