~/system-one-bench / compare/fastino-glide-vs-inception-mercury-decide
Head to head
GLiDE scores 84.1 and Mercury Decide 82.4 on the System One Index. GLiDE is more accurate on 13 of 16 text datasets, Mercury Decide on 2, and 1 are tied. Same cases, same request, same scorer.
| GLiDE | Mercury Decide | |
|---|---|---|
| System One Index | 84.1 | 82.4 |
| English | 83.0 | 83.7 |
| Arabic | 84.8 | 80.8 |
| Vision Index | 82.5 | 79.4 |
| Calibration error | 3.6% | 13.1% |
| Median latency | – | 276 ms |
| Cost per 1,000 calls | $0.017 | Free |
| GLiDE | Mercury Decide | |
|---|---|---|
| AG News English | 92.0 | 92.0 |
| SST-2 English | 96.0 | 95.5 |
| SST-5 English | 61.5 | 52.0 |
| DAIR Emotion English | 61.5 | 91.5 |
| Banking77 English | 91.5 | 86.0 |
| BoolQ English | 92.0 | 90.5 |
| XNLI English English | 94.7 | 91.3 |
| Typed decisions English | 75.0 | 71.1 |
| XNLI French French | 88.0 | 82.7 |
| XNLI Arabic Arabic | 83.3 | 76.0 |
| MASSIVE scenario Arabic | 73.5 | 72.0 |
| MASSIVE intent Arabic | 91.0 | 85.0 |
| OSACT4 offensive Arabic | 95.0 | 92.0 |
| Tweet sentiment Arabic | 73.5 | 62.5 |
| MIRACL rerank Arabic | 87.0 | 88.0 |
| MIRACL pairwise Arabic | 90.5 | 90.0 |
| GLiDE | Mercury Decide | |
|---|---|---|
| SST-2, as an image | 94.0 | 93.0 |
| Arabic sentiment, as an image | 51.5 | 39.0 |
| Oxford-IIIT Pet | 84.5 | 84.5 |
| VQAv2 yes/no | 89.0 | 87.0 |
| CountBench | 93.5 | 93.5 |
Model pages: GLiDE · Mercury Decide. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.