General Evaluation on LiveBench
46.83Accuracyphi-balancing
Evaluation Results
| Method | Links | |
|---|---|---|
| phi-balancingModel=Moonlight-16B-A3B-Instruct2026.05 | 46.83 | |
| ST-MoEModel=Moonlight-16B-A3B-Instruct2026.05 | 42.21 | |
| Frozen checkpointModel=Moonlight-16B-A3B-Instruct2026.05 | 19.57 | |
| phi-balancingModel=DeepSeek-V2-Lite2026.05 | 19.42 | |
| phi-balancingModel=DeepSeek-V2-Lite2026.05 | 19.34 | |
| ST-MoEModel=DeepSeek-V2-Lite2026.05 | 18.8 | |
| phi-balancingModel=DeepSeek-MoE-Chat2026.05 | 17.85 | |
| ST-MoEModel=DeepSeek-V2-Lite2026.05 | 16.85 | |
| ST-MoEModel=DeepSeek-MoE-Chat2026.05 | 14.62 | |
| phi-balancingModel=DeepSeek-MoE-Chat2026.05 | 14.15 | |
| Frozen checkpointModel=DeepSeek-V2-Lite2026.05 | 13.93 | |
| Frozen checkpointModel=DeepSeek-V2-Lite2026.05 | 13.93 | |
| ST-MoEModel=DeepSeek-MoE-Chat2026.05 | 13.79 | |
| Frozen checkpointModel=DeepSeek-MoE-Chat2026.05 | 5.92 | |
| Frozen checkpointModel=DeepSeek-MoE-Chat2026.05 | 5.92 |