Multi-task Language Understanding on MMLU-IT
81.5AccuracyQwen3-30B-A3B
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen3-30B-A3BGroup=Larger, Configuration=Standard-config2026.03 | 81.5 | |
| Qwen3-30B-A3BGroup=Larger, Configuration=Optimal-config2026.03 | 81.5 | |
| Gpt-oss-20b-highGroup=Larger, Configuration=Optimal-config2026.03 | 80.4 | |
| Gpt-oss-20b-highGroup=Larger, Configuration=Standard-config2026.03 | 80.2 | |
| Qwen3-14BGroup=Larger, Configuration=Standard-config2026.03 | 77.6 | |
| Qwen3-14BGroup=Larger, Configuration=Optimal-config2026.03 | 77.6 | |
| gemma-3-12b-itGroup=Larger, Configuration=Optimal-config2026.03 | 69 | |
| gemma-3-12b-itGroup=Larger, Configuration=Standard-config2026.03 | 68.5 | |
| gemma-2-9b-itGroup=Comparable, Configuration=Standard-config2026.03 | 67.2 | |
| gemma-2-9b-itGroup=Comparable, Configuration=Optimal-config2026.03 | 66.5 | |
| EngGPT2-16B-A3BGroup=Comparable, Configuration=Standard-config2026.03 | 65.5 | |
| EngGPT2-16B-A3BGroup=Comparable, Configuration=Optimal-config2026.03 | 65.5 | |
| Llama-3.1-8B-InstructGroup=Comparable, Configuration=Standard-config2026.03 | 60.6 | |
| Llama-3.1-8B-InstructGroup=Comparable, Configuration=Optimal-config2026.03 | 60.6 | |
| Moonlight-16B-A3B-InstructGroup=Comparable, Configuration=Standard-config2026.03 | 49.2 | |
| Moonlight-16B-A3B-InstructGroup=Comparable, Configuration=Optimal-config2026.03 | 49.2 |