Multiple Tasks on Foundational Benchmarks Average
59.8Average AccuracyMobileMoE-L
Evaluation results
| Method | Links | |
|---|---|---|
| MobileMoE-LActive Parameters=922M, Total Parameters=5.3B2026.05 | 59.8 | |
| MobileMoE-MActive Parameters=528M, Total Parameters=2.8B2026.05 | 55.4 | |
| Qwen3.5 2BActive Parameters=1.9B, Total Parameters=1.9B2026.05 | 53.3 | |
| OLMoE-1B-7BActive Parameters=1.3B, Total Parameters=6.9B2026.05 | 52.4 | |
| SmolLM2 1.7BActive Parameters=1.7B, Total Parameters=1.7B2026.05 | 50.7 | |
| OLMo 2 1BActive Parameters=1.5B, Total Parameters=1.5B2026.05 | 48.9 | |
| MobileLLM-ProActive Parameters=1.1B, Total Parameters=1.1B2026.05 | 47 | |
| MobileMoE-SActive Parameters=272M, Total Parameters=1.3B2026.05 | 46.5 | |
| Qwen3.5 0.8BActive Parameters=749M, Total Parameters=749M2026.05 | 46.4 | |
| Llama 3.2 1BActive Parameters=1.2B, Total Parameters=1.2B2026.05 | 42.5 | |
| Gemma 3 1BActive Parameters=1.0B, Total Parameters=1.0B2026.05 | 41.6 | |
| SmolLM2 360MActive Parameters=362M, Total Parameters=362M2026.05 | 39.3 | |
| Gemma 3 270MActive Parameters=270M, Total Parameters=270M2026.05 | 33.5 |