| code-generation-on-humaneval | MobileMoE-L (922M) | – | Pass@1: 65.2 |
| code-generation-on-mbpp | MobileMoE-L (922M) | – | Accuracy: 52.4 |
| common-sense-reasoning-on-arc-easy | MobileMoE-L (922M) | – | Accuracy: 81.9 |
| common-sense-reasoning-on-winogrande | MobileMoE-L (922M) | – | Accuracy: 66.1 |
| instruction-following-on-ifbench | MobileMoE-L (922M) | – | Accuracy: 20.1 |
| multiple-choice-question-answering-mcqa-on-gpqa-diamond | MobileMoE-L (922M) | – | Accuracy: 33.8 |
| question-answering-on-boolq | MobileMoE-L (922M) | – | Accuracy: 81.1 |
| question-answering-on-piqa | MobileMoE-L (922M) | – | Accuracy: 78.9 |
| sentence-completion-on-hellaswag | MobileMoE-L (922M) | – | Accuracy: 73.0 |
| social-commonsense-reasoning-on-siqa | MobileMoE-L (922M) | – | Accuracy: 53.4 |