OJBench
text
+
+
+
+
About
OJBench is a competition-level code benchmark comprising 232 challenging programming problems designed to assess large language models' code reasoning abilities in competitive programming contexts. It evaluates models' capacity to solve complex algorithmic problems, implement efficient solutions, and demonstrate advanced programming skills at the level required for programming competitions.
+
+
+
+
Evaluation Stats
Total Models4
Organizations2
Verified Results0
Self-Reported4
+
+
+
+
Benchmark Details
Max Score1
Language
en
+
+
+
+
Performance Overview
Score distribution and top performers
Score Distribution
4 models
Top Score
32.5%
Average Score
29.1%
High Performers (80%+)
0Top Organizations
#1Alibaba Cloud / Qwen Team
2 models
31.1%
#2Moonshot AI
2 models
27.1%
+
+
+
+
Leaderboard
4 models ranked by performance on OJBench
| License | Links | ||||
|---|---|---|---|---|---|
| Jul 25, 2025 | Apache 2.0 | 32.5% | |||
| Sep 10, 2025 | Apache 2.0 | 29.7% | |||
| Jul 11, 2025 | MIT | 27.1% | |||
| Sep 5, 2025 | MIT | 27.1% |
