system-one-bench

~/system-one-bench / compare/patrick-18-dexy-decision-9b-vs-perplexity-decider-v1-2

Head to head

Dexy-Decision-9B vs Decider V1.2

Dexy-Decision-9B scores 84.7 and Decider V1.2 84.6 on the System One Index. Dexy-Decision-9B is more accurate on 6 of 16 text datasets, Decider V1.2 on 7, and 3 are tied. Same cases, same request, same scorer.

Overall

Dexy-Decision-9BDecider V1.2
System One Index84.784.6
English82.382.0
Arabic86.886.9
Vision Index–83.8
Calibration error3.5%0.7%
Median latency–178 ms
Cost per 1,000 calls–$0.007

Text datasets, accuracy

Dexy-Decision-9BDecider V1.2
AG News English91.591.5
SST-2 English95.595.5
SST-5 English60.060.0
DAIR Emotion English54.060.0
Banking77 English93.088.0
BoolQ English89.591.5
XNLI English English94.794.0
Typed decisions English80.075.1
XNLI French French89.390.0
XNLI Arabic Arabic89.390.0
MASSIVE scenario Arabic89.579.0
MASSIVE intent Arabic91.591.0
OSACT4 offensive Arabic91.595.0
Tweet sentiment Arabic67.075.5
MIRACL rerank Arabic88.086.5
MIRACL pairwise Arabic91.091.5

Image datasets, accuracy

Dexy-Decision-9BDecider V1.2
SST-2, as an image–94.5
Arabic sentiment, as an image–57.0
Oxford-IIIT Pet–85.5
VQAv2 yes/no–86.5
CountBench–95.5

Model pages: Dexy-Decision-9B · Decider V1.2. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.