Multitask Language Understanding on MMLU Russian (test)
76AccuracyTrinity Large (MoE)
Evaluation Results
| Method | Links | |
|---|---|---|
| Trinity Large (MoE)Architecture=MoE2026.02 | 76 | |
| Qwen3 8BParameters=8B2026.02 | 69 | |
| Qwen3 4BParameters=4B2026.02 | 64 | |
| Llama-3.1 8BParameters=8B2026.02 | 53 | |
| Datology 8BParameters=8B2026.02 | 52 | |
| Granite-4.0 MicroModel Scale=Micro2026.02 | 50 | |
| SmolLM3 3BParameters=3B2026.02 | 48 | |
| Llama-3.2 3BParameters=3B2026.02 | 43 | |
| Datology 3BParameters=3B2026.02 | 41 | |
| Llama-3.2 1BParameters=1B2026.02 | 31 | |
| LFM2.5 1.2BParameters=1.2B2026.02 | 31 |