Code Generation on LiveBench (test)
56.5Sig. ScoreQwen2.5-14B-Ins. + CoSPlay w/ Cluster
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Qwen2.5-14B-Ins. + CoSPlay w/ ClusterBackbone=Ins., #GT Data=02026.05 | 56.5 | 70.1 | 47.9 | 57.6 | |
| Qwen2.5-14B-Ins. + CoSPlayBackbone=Ins., #GT Data=02026.05 | 54.9 | 70.1 | 47.9 | 55.7 | |
| CURE-14B + CoSPlay w/ ClusterBackbone=Ins., #GT Data=4.5k2026.05 | 54.9 | 71.5 | 52.4 | 60.7 | |
| CURE-7B + CoSPlay w/ ClusterBackbone=Ins., #GT Data=4.5k2026.05 | 53.6 | 66.2 | 45.4 | 52.3 | |
| CURE-14BBackbone=Ins., #GT Data=4.5k2026.05 | 53.6 | 77.3 | 47.7 | 59.9 | |
| CURE-14B + CoSPlayBackbone=Ins., #GT Data=4.5k2026.05 | 53.6 | 71.5 | 52.4 | 61.2 | |
| CURE-7BBackbone=Ins., #GT Data=4.5k2026.05 | 52.3 | 54.6 | 37.4 | 51.6 | |
| CURE-7B + CoSPlayBackbone=Ins., #GT Data=4.5k2026.05 | 51.3 | 66.2 | 45.4 | 52.1 | |
| Qwen2.5-7B-Ins. + CoSPlay w/ ClusterBackbone=Ins., #GT Data=02026.05 | 50.3 | 66.7 | 38.8 | 49.7 | |
| Qwen2.5-14B-Instruct + CodeTBackbone=Ins., #GT Data=-2026.05 | 49.5 | 26.4 | 39.2 | 52.1 | |
| Qwen2.5-14B-Instruct + ClusterBackbone=Ins., #GT Data=-2026.05 | 48.4 | 26.4 | 39.2 | 52.1 | |
| Qwen2.5-14B-Coder-Ins.Backbone=Coder-Ins., #GT Data=-2026.05 | 48.2 | 37.8 | 47.2 | 53.9 | |
| Qwen2.5-14B-InstructBackbone=Ins., #GT Data=-2026.05 | 47.7 | 26.4 | 39.2 | 51 | |
| Qwen2.5-7B-Ins. + CoSPlayBackbone=Ins., #GT Data=02026.05 | 47.4 | 66.7 | 38.8 | 50 | |
| DeepSeek-Coder-V2-16BBackbone=-, #GT Data=-2026.05 | 45.8 | 31 | 37.4 | 38.3 | |
| AZR-7B-CoderBackbone=Coder, #GT Data=02026.05 | 44.5 | 23.1 | 12.5 | 34.9 | |
| Seed-Coder-8B-InstructBackbone=-, #GT Data=-2026.05 | 43.5 | 26.9 | 37.6 | 45.1 | |
| AceCoder-7B-RuleBackbone=Coder-Ins., #GT Data=22k2026.05 | 43.5 | 17.2 | 36.1 | 45.8 | |
| AceCoder-7B-RMBackbone=Coder-Ins., #GT Data=329k2026.05 | 43 | 17.7 | 37.5 | 48.4 | |
| Qwen2.5-7B-Instruct + ClusterBackbone=Ins., #GT Data=-2026.05 | 42.7 | 9 | 33.1 | 42.7 | |
| Qwen2.5-7B-Coder-Ins.Backbone=Coder-Ins., #GT Data=-2026.05 | 42.7 | 12.5 | 36.3 | 47.7 | |
| Qwen2.5-14B-CoderBackbone=Coder, #GT Data=-2026.05 | 42.2 | 13.9 | 10.8 | 30.2 | |
| Qwen2.5-7B-CoderBackbone=Coder, #GT Data=-2026.05 | 40.6 | 12.1 | 3.8 | 19 | |
| AZR-14B-CoderBackbone=Coder, #GT Data=02026.05 | 39.8 | 22.2 | 38.9 | 43.5 | |
| Qwen2.5-7B-Instruct + CodeTBackbone=Ins., #GT Data=-2026.05 | 37.5 | 9 | 33.1 | 37.5 | |
| Qwen2.5-7B-InstructBackbone=Ins., #GT Data=-2026.05 | 36.2 | 9 | 33.1 | 36.2 |