Coding Reasoning on LeetCode-Contest
66.11Accuracy (%)COPT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| COPTBackbone=Qwen3-8B, Reasoning Effort=High (Peak Accuracy)2026.05 | 66.11 | 7,607 | |
| COPTBackbone=Qwen3-8B, Reasoning Effort=Token-saving / Matched Accuracy2026.05 | 61.11 | 6,993 | |
| CoTBackbone=Qwen3-8B2026.05 | 59.44 | 7,306 | |
| CoT (Greedy)Backbone=Qwen3-8B2026.05 | 57.22 | 6,975 |