system-one-bench

~/system-one-bench / compare/baseten-mammoth-1-vs-together-ai-tev1-4b-experimental

Head to head

Mammoth 1 vs Tev1 4B Experimental

Mammoth 1 scores 82.9 and Tev1 4B Experimental 71.5 on the System One Index. Mammoth 1 is more accurate on 15 of 16 text datasets, Tev1 4B Experimental on 1. Same cases, same request, same scorer.

Overall

Mammoth 1Tev1 4B Experimental
System One Index82.971.5
English82.566.4
Arabic82.875.8
Vision Indexโ€“โ€“
Calibration error4.9%5.3%
Median latencyโ€“232 ms
Cost per 1,000 calls$0.037$0.019

Text datasets, accuracy

Mammoth 1Tev1 4B Experimental
AG News English92.091.5
SST-2 English96.094.0
SST-5 English59.043.0
DAIR Emotion English59.056.5
Banking77 English94.51.0
BoolQ English92.091.0
XNLI English English91.387.3
Typed decisions English76.467.0
XNLI French French86.082.7
XNLI Arabic Arabic78.776.0
MASSIVE scenario Arabic75.056.5
MASSIVE intent Arabic91.070.0
OSACT4 offensive Arabic91.584.5
Tweet sentiment Arabic64.570.5
MIRACL rerank Arabic89.085.5
MIRACL pairwise Arabic90.087.5

Model pages: Mammoth 1 ยท Tev1 4B Experimental. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.