Code Generation on AtCoder
100Easy AccuracyMicroCoder-GRPO
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| MicroCoder-GRPOModel=Qwen3 4B, Training Context=4K, Testing Context=4K2026.03 | 100 | 49 | 9.6 | 39.7 | |
| GRPOModel=Qwen3 4B, Training Context=4K, Testing Context=8K2026.03 | 100 | 45.2 | 12.9 | 40.6 | |
| DAPOModel=Qwen3 4B, Training Context=4K, Testing Context=8K2026.03 | 100 | 43.3 | 11.3 | 39.3 | |
| MicroCoder-GRPOModel=Qwen3 4B, Training Context=4K, Testing Context=8K2026.03 | 100 | 48.1 | 12.9 | 41.3 | |
| GRPOModel=DS 8B, Training Context=16K, Testing Context=16K2026.03 | 100 | 39.4 | 5.8 | 35.5 | |
| DAPOModel=DS 8B, Training Context=16K, Testing Context=16K2026.03 | 100 | 44.2 | 8.3 | 37.9 | |
| MicroCoder-GRPOModel=DS 8B, Training Context=16K, Testing Context=16K2026.03 | 100 | 43.3 | 10 | 38.6 | |
| GRPOModel=Qwen3 4B, Training Context=4K, Testing Context=4K2026.03 | 98.1 | 46.2 | 11.7 | 39.7 | |
| DAPOModel=Qwen3 4B, Training Context=4K, Testing Context=4K2026.03 | 97.1 | 49 | 10 | 39.3 | |
| Qwen3-4B-InstructTesting Context=4K, RL Algorithm=None2026.03 | 95.2 | 44.2 | 9.2 | 37.3 | |
| Qwen3-4B-InstructTesting Context=8K, RL Algorithm=None2026.03 | 95.2 | 45.2 | 10 | 37.9 | |
| MicroCoder-GRPOModel=Qwen3 1.7B, Training Context=4K, Testing Context=4K2026.03 | 83.7 | 24 | 2.9 | 26.6 | |
| MicroCoder-GRPOModel=Qwen3 1.7B, Training Context=4K, Testing Context=8K2026.03 | 82.7 | 26.9 | 3.7 | 27.5 | |
| DAPOModel=Qwen3 1.7B, Training Context=4K, Testing Context=4K2026.03 | 76.9 | 15.4 | 2.9 | 23 | |
| DAPOModel=Qwen3 1.7B, Training Context=4K, Testing Context=8K2026.03 | 76.9 | 15.4 | 2.9 | 23 | |
| GRPOModel=Qwen3 1.7B, Training Context=4K, Testing Context=4K2026.03 | 70.2 | 20.2 | 1.7 | 21.9 | |
| GRPOModel=Qwen3 1.7B, Training Context=4K, Testing Context=8K2026.03 | 70.2 | 20.2 | 1.7 | 21.9 | |
| Qwen3-1.7B InstructTesting Context=4K, RL Algorithm=None2026.03 | 69.2 | 10.6 | 1.3 | 19.2 | |
| Qwen3-1.7B InstructTesting Context=8K, RL Algorithm=None2026.03 | 69.2 | 10.6 | 1.3 | 19.2 |