Code on CRUX
55.08Accuracy @5CoReward
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CoRewardBackbone=Qwen2.5-7B2026.04 | 55.08 | — | |
| JURY-RLBackbone=Qwen2.5-7B2026.04 | 54.37 | — | |
| GT-RewardBackbone=Qwen2.5-7B2026.04 | 53.67 | — | |
| LLM-KDBackbone=Qwen2.5-7B2026.04 | 53.57 | — | |
| Self-CertaintyBackbone=Qwen2.5-7B2026.04 | 53.18 | — | |
| Majority-VotingBackbone=Qwen2.5-7B2026.04 | 52.2 | — | |
| LLM-as-a-JudgeBackbone=Qwen2.5-7B2026.04 | 51.9 | — | |
| EntropyBackbone=Qwen2.5-7B2026.04 | 51.08 | — | |
| LLM-as-a-JudgeBackbone=Qwen3-1.7B-Base2026.04 | 34.98 | — | |
| CoRewardBackbone=Qwen3-1.7B-Base2026.04 | 34.1 | — | |
| JURY-RLBackbone=Qwen3-1.7B-Base2026.04 | 33.98 | — | |
| LLM-KDBackbone=Qwen3-1.7B-Base2026.04 | 33.8 | — | |
| GT-RewardBackbone=Qwen3-1.7B-Base2026.04 | 33.73 | — | |
| GT-RewardBackbone=Llama-3.2-3B-Instruct2026.04 | 32.48 | — | |
| Majority-VotingBackbone=Qwen3-1.7B-Base2026.04 | 32 | — | |
| JURY-RLBackbone=Llama-3.2-3B-Instruct2026.04 | 31.6 | — | |
| LLM-as-a-JudgeBackbone=Llama-3.2-3B-Instruct2026.04 | 31.45 | — | |
| LLM-KDBackbone=Llama-3.2-3B-Instruct2026.04 | 31.2 | — | |
| CoRewardBackbone=Llama-3.2-3B-Instruct2026.04 | 30.87 | — | |
| Majority-VotingBackbone=Llama-3.2-3B-Instruct2026.04 | 30.83 | — | |
| EntropyBackbone=Qwen3-1.7B-Base2026.04 | 30.38 | — | |
| Before RLBackbone=Qwen2.5-7B2026.04 | 27.38 | — | |
| EntropyBackbone=Llama-3.2-3B-Instruct2026.04 | 26.3 | — | |
| Before RLBackbone=Llama-3.2-3B-Instruct2026.04 | 24.7 | — | |
| Self-CertaintyBackbone=Llama-3.2-3B-Instruct2026.04 | 24.55 | — | |
| Self-CertaintyBackbone=Qwen3-1.7B-Base2026.04 | 18.73 | — | |
| Before RLBackbone=Qwen3-1.7B-Base2026.04 | 7.52 | — | |
| NPG-Muse-7BBackbone=Qwen2.5-7B-Instruct-1M2025.08 | — | 59 | |
| NPG-Muse-8BBackbone=Qwen3-8B-Base2025.08 | — | 65.3 | |
| Qwen2.5-14B-Instruct-1MSize=14B, Type=Instruct, Context Length=1M2025.08 | — | 66.4 | |
| Qwen2.5-7B-Instruct-1MSize=7B, Type=Instruct, Context Length=1M2025.08 | — | 55.6 | |
| Qwen3-14B-BaseSize=14B, Type=Base2025.08 | — | 63.9 | |
| Qwen3-8B-BaseSize=8B, Type=Base2025.08 | — | 54.1 |