System One Bench

Head to head

Solar Decide Flash vs d3 edge

Solar Decide Flash scores 63.6 and d3 edge 65.9 on the System One Index. Solar Decide Flash is more accurate on 11 of 16 text datasets, d3 edge on 5. Same cases, same request, same scorer.

Overall

Solar Decide Flashd3 edge
System One Index63.665.9
English61.369.1
Arabic66.562.2
Vision Index––
Calibration error14.4%18.7%
Median latency586 ms–
Cost per 1,000 calls$0.042–

Text datasets, accuracy

Solar Decide Flashd3 edge
AG News English86.080.0
SST-2 English90.580.5
SST-5 English44.539.5
DAIR Emotion English54.557.0
Banking77 English1.090.0
BoolQ English86.079.5
XNLI English English65.384.0
Typed decisions English62.542.3
XNLI French French61.366.0
XNLI Arabic Arabic52.059.3
MASSIVE scenario Arabic46.041.0
MASSIVE intent Arabic63.055.0
OSACT4 offensive Arabic84.572.5
Tweet sentiment Arabic65.561.5
MIRACL rerank Arabic71.570.0
MIRACL pairwise Arabic83.076.0

Model pages: Solar Decide Flash · d3 edge. Latency for API models is the median per call through OpenRouter, timed from one machine.