system-one-bench

~/system-one-bench / compare/fastino-glide-vs-nace-ai-drex-v2-0

Head to head

GLiDE vs Drex v2.0

GLiDE scores 84.1 and Drex v2.0 78.6 on the System One Index. GLiDE is more accurate on 14 of 16 text datasets, Drex v2.0 on 1, and 1 are tied. Same cases, same request, same scorer.

Overall

GLiDEDrex v2.0
System One Index84.178.6
English83.077.8
Arabic84.879.6
Vision Index82.5–
Calibration error3.6%1.5%
Median latency–294 ms
Cost per 1,000 calls$0.017$0.015

Text datasets, accuracy

GLiDEDrex v2.0
AG News English92.088.0
SST-2 English96.094.0
SST-5 English61.556.0
DAIR Emotion English61.555.5
Banking77 English91.580.5
BoolQ English92.092.5
XNLI English English94.787.3
Typed decisions English75.068.7
XNLI French French88.077.3
XNLI Arabic Arabic83.371.3
MASSIVE scenario Arabic73.563.5
MASSIVE intent Arabic91.082.5
OSACT4 offensive Arabic95.094.5
Tweet sentiment Arabic73.569.5
MIRACL rerank Arabic87.087.0
MIRACL pairwise Arabic90.589.0

Image datasets, accuracy

GLiDEDrex v2.0
SST-2, as an image94.0–
Arabic sentiment, as an image51.5–
Oxford-IIIT Pet84.5–
VQAv2 yes/no89.0–
CountBench93.5–

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