system-one-bench

~/system-one-bench / compare/meragpt-decider-1-vs-nace-ai-drex-v1-5

Head to head

Decider 1 vs Drex v1.5

Decider 1 scores 61.8 and Drex v1.5 77.6 on the System One Index. Decider 1 is more accurate on 4 of 16 text datasets, Drex v1.5 on 12. Same cases, same request, same scorer.

Overall

Decider 1Drex v1.5
System One Index61.877.6
English67.477.6
Arabic54.777.0
Vision Indexโ€“โ€“
Calibration error7.9%10.8%
Median latencyโ€“198 ms
Cost per 1,000 calls$0.012$0.010

Text datasets, accuracy

Decider 1Drex v1.5
AG News English87.590.0
SST-2 English93.094.5
SST-5 English55.553.5
DAIR Emotion English67.553.5
Banking77 English1.087.0
BoolQ English80.089.0
XNLI English English80.788.7
Typed decisions English74.264.8
XNLI French French67.382.0
XNLI Arabic Arabic67.374.0
MASSIVE scenario Arabic3.058.5
MASSIVE intent Arabic3.577.5
OSACT4 offensive Arabic86.090.5
Tweet sentiment Arabic64.063.5
MIRACL rerank Arabic81.089.0
MIRACL pairwise Arabic78.086.0

Model pages: Decider 1 ยท Drex v1.5. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.