General Capability on Aggregate (GPQA-D, GSM8K, HumanEval, MATH-500, MBPP, MMLU-Pro)
75.9Average AccuracyFOREVER
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FOREVERModel Scale=14B, Training Protocol=Capability-continual2026.05 | 75.9 | — | — | |
| FOREVERModel Scale=14B2026.05 | 75.9 | — | — | |
| FOREVERModel Scale=8B, Training Protocol=Capability-continual2026.05 | 75.6 | — | — | |
| MTLModel Scale=14B, Training Protocol=Joint-training2026.05 | 74.6 | — | — | |
| MTLModel Scale=14B2026.05 | 74.6 | — | — | |
| GCWM-DiamondScale=14B2026.05 | 74.32 | 1.39 | 0.8333 | |
| GCWMModel Scale=14B, Training Protocol=Capability-continual2026.05 | 74.3 | — | — | |
| GCWMModel Scale=14B2026.05 | 74.3 | — | — | |
| EWCModel Scale=14B, Training Protocol=Capability-continual2026.05 | 73.5 | — | — | |
| EWCModel Scale=14B2026.05 | 73.5 | — | — | |
| MTLModel Scale=8B, Training Protocol=Joint-training2026.05 | 73.3 | — | — | |
| OPCMScale=14B2026.05 | 72.93 | — | — | |
| OPCMModel Scale=14B, Training Protocol=Capability-continual2026.05 | 72.9 | — | — | |
| OPCMModel Scale=14B2026.05 | 72.9 | — | — | |
| EWCModel Scale=8B, Training Protocol=Capability-continual2026.05 | 72.6 | — | — | |
| GCWM-DiamondScale=8B2026.05 | 72.54 | 1.61 | 0.5 | |
| FOREVERModel Scale=4B, Training Protocol=Capability-continual2026.05 | 72.5 | — | — | |
| GCWMModel Scale=8B, Training Protocol=Capability-continual2026.05 | 72.5 | — | — | |
| AIMMergingModel Scale=14B, Training Protocol=Capability-continual2026.05 | 72.2 | — | — | |
| AIMMergingModel Scale=14B2026.05 | 72.2 | — | — | |
| Seq. SFTModel Scale=4B, Training Protocol=Capability-continual2026.05 | 71.6 | — | — | |
| L&SModel Scale=14B, Training Protocol=Capability-continual2026.05 | 71.3 | — | — | |
| L&SModel Scale=14B2026.05 | 71.3 | — | — | |
| OPCMScale=8B2026.05 | 70.93 | — | — | |
| OPCMModel Scale=8B, Training Protocol=Capability-continual2026.05 | 70.9 | — | — | |
| OPCMModel Scale=4B, Training Protocol=Capability-continual2026.05 | 70.5 | — | — | |
| OPCMScale=4B2026.05 | 70.45 | — | — | |
| Seq. SFTModel Scale=14B, Training Protocol=Capability-continual2026.05 | 70.4 | — | — | |
| Seq. SFTModel Scale=14B2026.05 | 70.4 | — | — | |
| GCWMModel Scale=4B, Training Protocol=Capability-continual2026.05 | 70.3 | — | — | |
| Seq. SFTModel Scale=8B, Training Protocol=Capability-continual2026.05 | 70.3 | — | — | |
| GCWM-DiamondScale=4B2026.05 | 70.27 | -0.18 | 1 | |
| AIMMergingModel Scale=8B, Training Protocol=Capability-continual2026.05 | 70.2 | — | — | |
| L&SModel Scale=8B, Training Protocol=Capability-continual2026.05 | 69.6 | — | — | |
| MTLModel Scale=4B, Training Protocol=Joint-training2026.05 | 68.7 | — | — | |
| EWCModel Scale=4B, Training Protocol=Capability-continual2026.05 | 68.6 | — | — | |
| AIMMergingModel Scale=4B, Training Protocol=Capability-continual2026.05 | 65.6 | — | — | |
| L&SModel Scale=4B, Training Protocol=Capability-continual2026.05 | 64.7 | — | — | |
| FOREVERModel Scale=1.7B, Training Protocol=Capability-continual2026.05 | 58.3 | — | — | |
| GCWMModel Scale=1.7B, Training Protocol=Capability-continual2026.05 | 58.3 | — | — | |
| FOREVERModel Scale=1.7B2026.05 | 58.3 | — | — | |
| GCWMModel Scale=1.7B2026.05 | 58.3 | — | — | |
| GCWM-DiamondScale=1.7B2026.05 | 58.26 | 1.44 | 1 | |
| MTLModel Scale=1.7B, Training Protocol=Joint-training2026.05 | 57.1 | — | — | |
| MTLModel Scale=1.7B2026.05 | 57.1 | — | — | |
| OPCMScale=1.7B2026.05 | 56.82 | — | — | |
| OPCMModel Scale=1.7B, Training Protocol=Capability-continual2026.05 | 56.8 | — | — | |
| OPCMModel Scale=1.7B2026.05 | 56.8 | — | — | |
| EWCModel Scale=1.7B, Training Protocol=Capability-continual2026.05 | 54.7 | — | — | |
| EWCModel Scale=1.7B2026.05 | 54.7 | — | — | |
| AIMMergingModel Scale=1.7B, Training Protocol=Capability-continual2026.05 | 53.4 | — | — | |
| AIMMergingModel Scale=1.7B2026.05 | 53.4 | — | — | |
| L&SModel Scale=1.7B, Training Protocol=Capability-continual2026.05 | 52.4 | — | — | |
| L&SModel Scale=1.7B2026.05 | 52.4 | — | — | |
| Seq. SFTModel Scale=1.7B, Training Protocol=Capability-continual2026.05 | 51.9 | — | — | |
| Seq. SFTModel Scale=1.7B2026.05 | 51.9 | — | — | |
| MTLModel Scale=0.6B, Training Protocol=Joint-training2026.05 | 40.7 | — | — | |
| AIMMergingModel Scale=0.6B, Training Protocol=Capability-continual2026.05 | 40.1 | — | — | |
| AIMMergingScale=0.6B2026.05 | 40.07 | — | — | |
| FOREVERModel Scale=0.6B, Training Protocol=Capability-continual2026.05 | 40 | — | — | |
| GCWMModel Scale=0.6B, Training Protocol=Capability-continual2026.05 | 40 | — | — | |
| GCWM-DiamondScale=0.6B2026.05 | 39.96 | -0.11 | 0.5 | |
| L&SModel Scale=0.6B, Training Protocol=Capability-continual2026.05 | 39.6 | — | — | |
| EWCModel Scale=0.6B, Training Protocol=Capability-continual2026.05 | 38.9 | — | — | |
| OPCMModel Scale=0.6B, Training Protocol=Capability-continual2026.05 | 38.8 | — | — | |
| Seq. SFTModel Scale=0.6B, Training Protocol=Capability-continual2026.05 | 37.4 | — | — |