Multipl-E HumanEval
Multilingual
text
+
+
+
+
About
MultiPL-E HumanEval is a variant of the MultiPL-E benchmark specifically based on the HumanEval dataset, extending it to multiple programming languages. It evaluates code generation models' ability to solve the same programming problems across different languages, testing cross-language code generation consistency and programming language understanding.
+
+
+
+
Evaluation Stats
Total Models3
Organizations1
Verified Results0
Self-Reported3
+
+
+
+
Benchmark Details
Max Score1
Language
en
+
+
+
+
Performance Overview
Score distribution and top performers
Score Distribution
3 models
Top Score
75.2%
Average Score
63.8%
High Performers (80%+)
0Top Organizations
#1Meta
3 models
63.8%
+
+
+
+
Leaderboard
3 models ranked by performance on Multipl-E HumanEval
| License | Links | ||||
|---|---|---|---|---|---|
| Jul 23, 2024 | Llama 3.1 Community License | 75.2% | |||
| Jul 23, 2024 | Llama 3.1 Community License | 65.5% | |||
| Jul 23, 2024 | Llama 3.1 Community License | 50.8% |
