Computer Science Problem Solving on ALE-bench
661.64Average Score @5FrontierSmith (200)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| FrontierSmith (200)Model=Qwen3.5-27B2026.05 | 661.64 | 938.1 | |
| ALE-bench (40)Model=Qwen3.5-9B2026.05 | 657.4 | 750.15 | |
| FrontierSmith (200)Model=Qwen3.5-9B2026.05 | 633.58 | 782.3 | |
| FrontierCS (172)Model=Qwen3.5-9B2026.05 | 558.49 | 762.28 | |
| FrontierCS (172)Model=Qwen3.5-27B2026.05 | 543.8 | 843.8 | |
| HardTests (200)Model=Qwen3.5-27B2026.05 | 529.12 | 922.5 | |
| HardTests (200)Model=Qwen3.5-9B2026.05 | 397.18 | 512.83 | |
| Random Reward (172)Model=Qwen3.5-9B2026.05 | 376.82 | 466.73 | |
| BaseModel=Qwen3.5-27B2026.05 | 352.52 | 498 | |
| BaseModel=Qwen3.5-9B2026.05 | 327.22 | 327.22 |