Stability AI·released 2023-07-20
Scored on 7 benchmarks. 0% of its scores are independently reproduced — the rest are self-reported or unverified.
Every score, ranked, colored by evidence status.
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| ARC AI2 | 81.47accuracy (%) | self-reported· optimized | 2023-07-20 |
| HellaSwag | 78.8accuracy (%) | self-reported· optimized | 2023-07-20 |
| LAMBADA | 71.3accuracy (%) | self-reported· optimized | 2023-07-20 |
| GSM8K | 69.6accuracy (%) | self-reported· optimized | 2023-07-20 |
| PIQA | 66.6accuracy (%) | self-reported· optimized | 2023-07-20 |
| BBH | 59.07accuracy (%) | self-reported· optimized | 2023-07-20 |
| MMLU | 58.13accuracy (%) | self-reported· optimized | 2023-07-20 |