system-one-bench

~/system-one-bench / compare/fastino-glide-vs-openai-gpt-6-luna-decisions

Head to head

GLiDE vs GPT-6 Luna Decisions

GLiDE scores 84.1 and GPT-6 Luna Decisions 76.0 on the System One Index. GLiDE is more accurate on 14 of 16 text datasets, GPT-6 Luna Decisions on 1, and 1 are tied. Same cases, same request, same scorer.

Overall

GLiDEGPT-6 Luna Decisions
System One Index84.176.0
English83.075.2
Arabic84.878.8
Vision Index82.581.6
Calibration error3.6%9.8%
Median latency–215 ms
Cost per 1,000 calls$0.017$0.040

Text datasets, accuracy

GLiDEGPT-6 Luna Decisions
AG News English92.087.5
SST-2 English96.093.5
SST-5 English61.561.5
DAIR Emotion English61.558.5
Banking77 English91.577.5
BoolQ English92.085.5
XNLI English English94.766.7
Typed decisions English75.071.0
XNLI French French88.062.0
XNLI Arabic Arabic83.354.0
MASSIVE scenario Arabic73.566.0
MASSIVE intent Arabic91.084.5
OSACT4 offensive Arabic95.093.5
Tweet sentiment Arabic73.580.0
MIRACL rerank Arabic87.086.0
MIRACL pairwise Arabic90.587.5

Image datasets, accuracy

GLiDEGPT-6 Luna Decisions
SST-2, as an image94.088.5
Arabic sentiment, as an image51.567.0
Oxford-IIIT Pet84.587.0
VQAv2 yes/no89.080.5
CountBench93.585.0

Model pages: GLiDE · GPT-6 Luna Decisions. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.