Benchmark
LiveCodeBench
reasoning
general
code
text
LiveCodeBench is a holistic and contamination-free evaluation benchmark for large language models for code. It continuously collects new problems from programming contests (LeetCode, AtCoder, CodeForces) and evaluates four different scenarios: code generation, self-repair, code execution, and test output prediction. Problems are annotated with release dates to enable evaluation on unseen problems released after a model's training cutoff.
Top 10
榜单概览
数据说明
分数为 0-100 百分制综合成绩（Elo 榜为相对评分，仅供同榜内比较）。评测分数属事实数据，整理自公开评测体系与厂商公开资料，部分分数为厂商自报口径，不同榜单的评测方法与数据集可能存在差异，请结合多个榜单交叉参考。
Ranking
模型排名
| 名次 | 模型 | 机构 | 分数 | 来源 |
|---|---|---|---|---|
| 1 | Fugu Ultra | Sakana AI | 93.2 | 来源 |
| 2 | Fugu | Sakana AI | 92.9 | 来源 |
| 3 | Nemotron 3 Ultra 550B A55B | NVIDIA | 89.0 | 来源 |
| 4 | Grok 4 Fast | xAI | 80.0 | 来源 |
| 5 | DeepSeek-V3.2-Exp | DeepSeek | 74.1 | 来源 |
| 6 | GLM-4.5 | Zhipu AI | 72.9 | 来源 |
| 7 | Nemotron Nano 9B v2 | NVIDIA | 71.1 | 来源 |
| 8 | GLM-4.5-Air | Zhipu AI | 70.7 | 来源 |
| 9 | Kimi K2-Instruct-0905 | Moonshot AI | 53.7 | 来源 |