Russian-focused evaluation on MERA
38.1MERA ScoreGamayun
Evaluation results
| Method | Links | |
|---|---|---|
| GamayunNumber of Parameters=1.5B2025.12 | 38.1 | |
| QVikhrBase Model=Qwen32025.12 | 36.8 | |
| Qwen3Number of Parameters=1.7B2025.12 | 36.4 | |
| Qwen2.5Number of Parameters=1.5B2025.12 | 35.7 | |
| meno-tiny-0.1Base Model=Qwen32025.12 | 35.2 | |
| ruadaptBase Model=Qwen2.52025.12 | 33.7 | |
| VikhrLlamaBase Model=Llama3.22025.12 | 30 | |
| Llama3.2Number of Parameters=1.24B2025.12 | 29.8 |