system-one-bench

~/system-one-bench / compare/baseten-mammoth-1-vs-nace-ai-drex-v2-0

Head to head

Mammoth 1 vs Drex v2.0

Mammoth 1 scores 82.9 and Drex v2.0 78.6 on the System One Index. Mammoth 1 is more accurate on 13 of 16 text datasets, Drex v2.0 on 3. Same cases, same request, same scorer.

Overall

Mammoth 1Drex v2.0
System One Index82.978.6
English82.577.8
Arabic82.879.6
Vision Indexโ€“โ€“
Calibration error4.9%1.5%
Median latencyโ€“294 ms
Cost per 1,000 calls$0.037$0.015

Text datasets, accuracy

Mammoth 1Drex v2.0
AG News English92.088.0
SST-2 English96.094.0
SST-5 English59.056.0
DAIR Emotion English59.055.5
Banking77 English94.580.5
BoolQ English92.092.5
XNLI English English91.387.3
Typed decisions English76.468.7
XNLI French French86.077.3
XNLI Arabic Arabic78.771.3
MASSIVE scenario Arabic75.063.5
MASSIVE intent Arabic91.082.5
OSACT4 offensive Arabic91.594.5
Tweet sentiment Arabic64.569.5
MIRACL rerank Arabic89.087.0
MIRACL pairwise Arabic90.089.0

Model pages: Mammoth 1 ยท Drex v2.0. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.