Question Answering on HellaSwag, WinoGrande, LAMBADA, PIQA, OpenBookQA, ARC-E, and ARC-C
79.2HellaSwag AccuracyFP16
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| FP16Model=LLaMA3.1-8B-Instruct2026.04 | 79.2 | 74.1 | 73.2 | 80.9 | 43 | 79.6 | 55.2 | 69.3 | |
| FP16Model=LLaMA3-8B2026.04 | 79.1 | 72.9 | 75.5 | 80.7 | 44.6 | 77.6 | 53.5 | 69.1 | |
| DuQuant++Model=LLaMA3.1-8B-Instruct, Format=MXFP42026.04 | 77.3 | 70.6 | 70.2 | 79.3 | 42 | 78 | 52.1 | 67.1 | |
| DuQuant++*Model=LLaMA3.1-8B-Instruct, Format=MXFP42026.04 | 77 | 71.7 | 72.3 | 79.5 | 42.4 | 77 | 51.7 | 67.4 | |
| DuQuant++Model=LLaMA3-8B, Format=MXFP42026.04 | 76.6 | 72.2 | 73 | 78.7 | 41.8 | 74.7 | 48.2 | 66.5 | |
| DuQuant++*Model=LLaMA3-8B, Format=MXFP42026.04 | 76.6 | 71.7 | 74 | 79.5 | 42.6 | 75.3 | 50.1 | 67.1 | |
| MR-GPTQModel=LLaMA3-8B, Format=MXFP42026.04 | 76.4 | 69.7 | 71.8 | 77.8 | 43.4 | 74.5 | 49.4 | 66.1 | |
| MR-GPTQModel=LLaMA3.1-8B-Instruct, Format=MXFP42026.04 | 76.2 | 70.6 | 70.4 | 78.4 | 43.4 | 77.2 | 52.6 | 67 | |
| FP16Model=LLaMA3-8B-Instruct2026.04 | 75.7 | 71.4 | 71.5 | 78.6 | 42.8 | 78.8 | 55.6 | 67.8 | |
| FlatQuantModel=LLaMA3.1-8B-Instruct, Format=MXFP42026.04 | 75.7 | 71.1 | 69.6 | 78.7 | 41.8 | 78.3 | 52.1 | 66.8 | |
| FlatQuantModel=LLaMA3-8B, Format=MXFP42026.04 | 75.4 | 71.9 | 68.6 | 78.9 | 42.6 | 74 | 47.2 | 65.5 | |
| FP16Model=LLaMA3.2-3B2026.04 | 74 | 69.5 | 69.6 | 77.7 | 40.4 | 71.8 | 46.3 | 64.2 | |
| DuQuant++Model=LLaMA3-8B-Instruct, Format=MXFP42026.04 | 73.8 | 72.1 | 69.7 | 76.1 | 41.2 | 75.7 | 52.8 | 65.9 | |
| DuQuant++*Model=LLaMA3-8B-Instruct, Format=MXFP42026.04 | 73.7 | 71.6 | 70.3 | 76.7 | 41.2 | 77.2 | 50.7 | 65.9 | |
| MR-GPTQModel=LLaMA3-8B-Instruct, Format=MXFP42026.04 | 73.2 | 69 | 68.3 | 76 | 43 | 75.8 | 51.3 | 65.2 | |
| FlatQuantModel=LLaMA3-8B-Instruct, Format=MXFP42026.04 | 73 | 71.1 | 67.3 | 76.6 | 41.2 | 75.8 | 49.4 | 64.9 | |
| QuaRotModel=LLaMA3.1-8B-Instruct, Format=MXFP42026.04 | 72.4 | 68.4 | 62 | 76 | 39.8 | 71.8 | 46.3 | 62.4 | |
| QuaRot*Model=LLaMA3-8B, Format=MXFP42026.04 | 72.3 | 68.5 | 74.6 | 76.8 | 43.2 | 75 | 47.3 | 65.4 | |
| QuaRot*Model=LLaMA3.1-8B-Instruct, Format=MXFP42026.04 | 71.8 | 70.7 | 72.5 | 76.9 | 39.4 | 76.3 | 51.9 | 65.6 | |
| DuQuant++*Model=LLaMA3.2-3B, Format=MXFP42026.04 | 70.6 | 65.8 | 67.8 | 75 | 42.2 | 68.5 | 42.5 | 61.8 | |
| QuaRotModel=LLaMA3-8B, Format=MXFP42026.04 | 70.4 | 70 | 68.1 | 76.1 | 41.6 | 69.8 | 44 | 62.9 | |
| DuQuant++Model=LLaMA3.2-3B, Format=MXFP42026.04 | 70.3 | 65.1 | 65.3 | 75.4 | 39.2 | 67.9 | 42.9 | 60.9 | |
| MR-GPTQModel=LLaMA3.2-3B, Format=MXFP42026.04 | 70.1 | 65 | 66.6 | 76.1 | 37.8 | 70 | 42.8 | 61.2 | |
| QuaRot*Model=LLaMA3-8B-Instruct, Format=MXFP42026.04 | 69.2 | 68.8 | 71.4 | 75.4 | 39.6 | 75.8 | 48.2 | 64.1 | |
| QuaRotModel=LLaMA3-8B-Instruct, Format=MXFP42026.04 | 68 | 67.3 | 67.4 | 75.5 | 40.2 | 72.6 | 45.9 | 62.4 | |
| FlatQuantModel=LLaMA3.2-3B, Format=MXFP42026.04 | 67.9 | 65.1 | 61.6 | 74.3 | 37.2 | 66.5 | 40.4 | 59 | |
| QuaRot*Model=LLaMA3.2-3B, Format=MXFP42026.04 | 66.1 | 63.7 | 62.2 | 73.3 | 37.6 | 62.2 | 36.6 | 57.4 | |
| QuaRotModel=LLaMA3.2-3B, Format=MXFP42026.04 | 57.9 | 59.8 | 50.8 | 70.3 | 32.6 | 59.2 | 34.8 | 52.2 |