system-one-bench

~/system-one-bench / compare/baseten-joey-1-vs-google-gemini-decision-1

Head to head

Joey 1 vs Gemini Decision 1

Joey 1 scores 73.3 and Gemini Decision 1 81.6 on the System One Index. Joey 1 is more accurate on 1 of 16 text datasets, Gemini Decision 1 on 15. Same cases, same request, same scorer.

Overall

Joey 1Gemini Decision 1
System One Index73.381.6
English72.979.2
Arabic73.584.5
Vision Index–89.2
Calibration error2.1%3.6%
Median latency––
Cost per 1,000 calls$0.004$0.036

Text datasets, accuracy

Joey 1Gemini Decision 1
AG News English85.088.5
SST-2 English86.096.5
SST-5 English40.559.0
DAIR Emotion English52.559.5
Banking77 English93.079.5
BoolQ English80.089.0
XNLI English English84.087.3
Typed decisions English62.474.0
XNLI French French74.781.3
XNLI Arabic Arabic64.778.0
MASSIVE scenario Arabic71.572.5
MASSIVE intent Arabic84.587.0
OSACT4 offensive Arabic83.095.5
Tweet sentiment Arabic66.078.0
MIRACL rerank Arabic62.089.5
MIRACL pairwise Arabic83.091.0

Image datasets, accuracy

Joey 1Gemini Decision 1
SST-2, as an image–94.5
Arabic sentiment, as an image–75.0
Oxford-IIIT Pet–98.0
VQAv2 yes/no–86.0
CountBench–92.5

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