Code Generation on LiveCodeBench (Signal, UT, Code, BoN)
46.7Signal Score+ Cluster
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| + ClusterModel=DeepSeek-V3.2-685B, Model Category=Base Models2026.05 | 46.7 | 83.2 | 78.4 | 80 | |
| CoSPlay (Ours)Model=DeepSeek-V3.2-685B, Model Category=Base Models2026.05 | 46 | 83.2 | 78.4 | 79.3 | |
| BaselineModel=DeepSeek-V3.2-685B, Model Category=Base Models2026.05 | 45.3 | 85.1 | 75.7 | 82.7 | |
| + ClusterModel=AceCoder-Rule-7B, Model Category=RL-tuned Models2026.05 | 44.7 | 69.7 | 40 | 51.3 | |
| + ClusterModel=AceCoder-RM-7B, Model Category=RL-tuned Models2026.05 | 44 | 70.7 | 39.8 | 50 | |
| CoSPlay (Ours)Model=DeepSeek-Coder-V2-Lite-Instruct-16B, Model Category=Base Models2026.05 | 43.3 | 86.6 | 43.5 | 48 | |
| + ClusterModel=DeepSeek-Coder-V2-Lite-Instruct-16B, Model Category=Base Models2026.05 | 43.3 | 86.6 | 43.5 | 48.7 | |
| + ClusterModel=Gemini-2.0-Flash, Model Category=Base Models2026.05 | 43.3 | 83.5 | 61.5 | 65.3 | |
| CoSPlay (Ours)Model=AceCoder-Rule-7B, Model Category=RL-tuned Models2026.05 | 43.3 | 69.7 | 40 | 51.3 | |
| CoSPlay (Ours)Model=AceCoder-RM-7B, Model Category=RL-tuned Models2026.05 | 43.3 | 70.7 | 39.8 | 47.3 | |
| CoSPlay (Ours)Model=Seed-Coder-8B-Instruct, Model Category=Base Models2026.05 | 42.7 | 75.9 | 42.2 | 52.7 | |
| + ClusterModel=Seed-Coder-8B-Instruct, Model Category=Base Models2026.05 | 42.7 | 75.9 | 42.2 | 52.7 | |
| CoSPlay (Ours)Model=Gemini-2.0-Flash, Model Category=Base Models2026.05 | 42.7 | 83.5 | 61.5 | 63.3 | |
| BaselineModel=Gemini-2.0-Flash, Model Category=Base Models2026.05 | 42 | 50.2 | 56.6 | 62 | |
| BaselineModel=DeepSeek-Coder-V2-Lite-Instruct-16B, Model Category=Base Models2026.05 | 38 | 39.9 | 28.4 | 34.7 | |
| + ClusterModel=Absolute-Zero-7B, Model Category=RL-tuned Models2026.05 | 36 | 84.9 | 28.1 | 36.7 | |
| CoSPlay (Ours)Model=Absolute-Zero-7B, Model Category=RL-tuned Models2026.05 | 34.7 | 84.9 | 28.1 | 33.3 | |
| BaselineModel=Absolute-Zero-7B, Model Category=RL-tuned Models2026.05 | 32.7 | 31 | 11.6 | 24.7 | |
| BaselineModel=Seed-Coder-8B-Instruct, Model Category=Base Models2026.05 | 32 | 26.5 | 32.4 | 38.7 | |
| BaselineModel=AceCoder-Rule-7B, Model Category=RL-tuned Models2026.05 | 29.3 | 15.9 | 34 | 36 | |
| BaselineModel=AceCoder-RM-7B, Model Category=RL-tuned Models2026.05 | 29.3 | 17.2 | 34 | 38 |