Falcon 7B — Benchmarks
Benchmark scores for Falcon 7B aggregated from public leaderboards, with how it ranks among open models. See hardware requirements for what you need to run it.
Overall rank: #91 of 123 open modelscomposite 34/100 across 10 benchmarks in 3 categories · methodology
Knowledge
| Benchmark | Score | Open rank | All models |
|---|---|---|---|
| HellaSwag | 78.1% | #17 / 42 | #34 / 76 |
| ARC Challenge | 47.9% | #31 / 51 | #49 / 77 |
| BoolQ | 75.3% | #21 / 33 | #51 / 77 |
| OpenBookQA | 51.6% | #14 / 21 | #26 / 42 |
| TriviaQA | 64.6% | #14 / 18 | #33 / 39 |
| MMLU | 35.0% | #73 / 77 | #127 / 136 |
Math
| Benchmark | Score | Open rank | All models |
|---|---|---|---|
| GSM8K | 6.8% | #59 / 60 | #84 / 93 |
Reasoning
| Benchmark | Score | Open rank | All models |
|---|---|---|---|
| BIG-Bench Hard | 28.8% | #36 / 37 | #49 / 50 |
| WinoGrande | 67.2% | #28 / 46 | #56 / 80 |
| PIQA | 80.3% | #19 / 35 | #36 / 60 |
Scores aggregated from public benchmark sources (each linked from the benchmark pages). llmrun does not run these benchmarks.