Code Generation on LiveCodeBench Aug 2024–Feb 2025 v5
98.6Easy Success RateQwen3-30B-Thinking
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Qwen3-30B-Thinkingtype=Base Model2026.05 | 98.6 | 84.4 | 49.2 | 72.3 | |
| DuSTBase Model=Qwen3-30B-Thinking2026.05 | 98.6 | 87 | 54.1 | 75.2 | |
| DuSTBase Model=Qwen3-4B-Thinking2026.05 | 97.9 | 77.8 | 36.9 | 64.6 | |
| OpenReasoning-Nemotron-32B2026.05 | 97.6 | 82.8 | 53 | 73.2 | |
| DuSTBase Model=GPT-OSS-20B2026.05 | 97.5 | 81.4 | 53.1 | 72.8 | |
| OpenReasoning-Nemotron-7B2026.05 | 97 | 78 | 43.2 | 67.3 | |
| GPT-OSS-20Btype=Base Model2026.05 | 97 | 79.3 | 42.9 | 67.5 | |
| DeepSeek-R1-0528-Qwen3-8B2026.05 | 96.5 | 72.6 | 30.1 | 59.6 | |
| Klear-Reasoner-8B2026.05 | 96.4 | 77.3 | 35.1 | 63.3 | |
| Qwen3-4B-Thinkingtype=Base Model2026.05 | 95.9 | 73.3 | 36.1 | 62.3 | |
| OpenReasoning-Nemotron-14B2026.05 | 95.7 | 83.4 | 46.8 | 70.2 | |
| OpenCodeReasoning-1.1-32B2026.05 | 95.4 | 83.4 | 49.4 | 71.3 | |
| OpenCodeReasoning-1.1-14B2026.05 | 95.4 | 76.6 | 43.3 | 66.6 | |
| OpenCodeReasoning-1.1-7B2026.05 | 94.6 | 60.8 | 17.2 | 49.7 | |
| Qwen3-30B-Instructtype=Base Model2026.05 | 92.4 | 47.4 | 19.3 | 46 | |
| DuSTBase Model=Qwen3-30B-Instruct2026.05 | 90.2 | 53.5 | 18.7 | 47.1 | |
| TinyR1-32B2026.05 | 88.1 | 85.6 | 51.1 | 71.2 | |
| Skywork-OR1-7B2026.05 | 88.1 | 52.2 | 20 | 46.8 |