Knowledge & Reasoning on MMLU-Redux
79.8AccuracySigma-MoE-Tiny
Evaluation results
| Method | Links | |
|---|---|---|
| Sigma-MoE-TinyArchitecture=MoE, # Activated Params=0.5B, # Total Params=20B2025.12 | 79.8 | |
| Phi-3.5-MoEArchitecture=MoE, # Activated Params=6.6B, # Total Params=42B2025.12 | 78.6 | |
| Qwen3-1.7BArchitecture=Dense, # Activated Params=1.7B, # Total Params=1.7B2025.12 | 73.9 | |
| DeepSeek-R1-Distill-Qwen-7BArchitecture=Dense, # Activated Params=7B, # Total Params=7B2025.12 | 68.5 | |
| DeepSeek-R1-Distill-Llama-8BArchitecture=Dense, # Activated Params=8B, # Total Params=8B2025.12 | 66.4 | |
| Engram-40BShots=5-shot2026.01 | 64.5 | |
| Engram-27BShots=5-shot2026.01 | 64 | |
| MoE-27BShots=5-shot2026.01 | 60.6 | |
| Dense-4BShots=5-shot2026.01 | 50.7 |