MultiPL-E MBPP

Overview

MultiPL-E MBPP is a multilingual code generation benchmark derived from MBPP (Mostly Basic Python Programming). It extends the original MBPP to 18 programming languages, enabling cross-lingual evaluation of code generation capabilities.

Task Description

Task Type: Multilingual Code Generation
Input: Programming problem prompt with docstring
Output: Complete code solution that passes test cases
Languages: 18 languages (C++, TypeScript, Shell, C#, Go, Java, Lua, JavaScript, PHP, Perl, Racket, R, Rust, Scala, Swift, Ruby, D, Julia)

Key Features

Multilingual evaluation across 18 programming languages
Execution-based evaluation with test cases
Supports pass@k metric for code generation
Docker sandbox environment for safe code execution
Derived from MBPP with consistent problem difficulty

Evaluation Notes

Sandbox Required: Requires sandbox environment for safe code execution
Default evaluation uses test split
Primary metric: Accuracy with pass@k aggregation
Timeout: 30 seconds per test case
See sandbox documentation for setup

Properties

Property	Value
Benchmark Name	`multiple_mbpp`
Dataset ID	evalscope/MultiPL-E
Paper	N/A
Tags	`Coding`
Metrics	`acc`
Default Shots	0-shot
Evaluation Split	`test`
Aggregation	`mean_and_pass_at_k`

Data Statistics

Metric	Value
Total Samples	6,987
Prompt Length (Mean)	379.32 chars
Prompt Length (Min/Max)	266 / 1002 chars

Per-Subset Statistics:

Subset	Samples	Prompt Mean	Prompt Min	Prompt Max
`mbpp-cpp`	397	407.55	317	1002
`mbpp-ts`	390	359.57	295	674
`mbpp-sh`	382	362.53	296	696
`mbpp-cs`	386	549.09	481	867
`mbpp-go`	374	402.21	332	723
`mbpp-java`	386	553	474	889
`mbpp-lua`	397	332.52	277	651
`mbpp-js`	397	325.48	270	645
`mbpp-php`	397	336.17	280	655
`mbpp-pl`	396	337.87	282	657
`mbpp-rkt`	397	342.25	287	660
`mbpp-r`	397	326.42	271	644
`mbpp-rs`	354	351.9	285	669
`mbpp-scala`	396	434.01	364	750
`mbpp-swift`	396	352.64	285	667
`mbpp-rb`	397	321.34	266	641
`mbpp-d`	358	376.49	308	693
`mbpp-jl`	390	363.01	292	686

Sample Example

Subset: mbpp-cpp

{
  "input": [
    {
      "id": "209e8d38",
      "content": "```cpp\n#include<assert.h>\n#include<bits/stdc++.h>\n// Write a cppthon function to identify non-prime numbers.\nbool is_not_prime(long n) {\n\n```\n\nPlease complete the above code according to the requirements in the docstring. Write the complete code and wrap it in markdown fenced code. The code should not contain `Main` function."
    }
  ],
  "target": "",
  "id": 0,
  "group_id": 0,
  "metadata": {
    "tests": "}\nint main() {\n    auto candidate = is_not_prime;\n    assert(candidate((2)) == (false));\n    assert(candidate((10)) == (true));\n    assert(candidate((35)) == (true));\n    assert(candidate((37)) == (false));\n}\n",
    "stop_tokens": [
      "\n}"
    ],
    "task_id": "mbpp_3_is_not_prime",
    "language": "cpp",
    "doctests": "transform"
  }
}

Prompt Template

Prompt Template:

{prompt}

Sandbox Configuration

This benchmark requires a sandbox environment for code execution.

{
  "image": "volcengine/sandbox-fusion:server-20250609",
  "tools_config": {
    "shell_executor": {},
    "python_executor": {},
    "multi_code_executor": {}
  },
  "memory_limit": "2g",
  "cpu_limit": "2.0"
}

Usage

Using CLI

evalscope eval \
    --model YOUR_MODEL \
    --api-url OPENAI_API_COMPAT_URL \
    --api-key EMPTY_TOKEN \
    --datasets multiple_mbpp \
    --sandbox '{"enabled": true}' \
    --limit 10  # Remove this line for formal evaluation

Using Python

from evalscope import run_task
from evalscope.config import TaskConfig

task_cfg = TaskConfig(
    model='YOUR_MODEL',
    api_url='OPENAI_API_COMPAT_URL',
    api_key='EMPTY_TOKEN',
    datasets=['multiple_mbpp'],
    sandbox={'enabled': True},
    dataset_args={
        'multiple_mbpp': {
            # subset_list: ['mbpp-cpp', 'mbpp-ts', 'mbpp-sh']  # optional, evaluate specific subsets
        }
    },
    limit=10,  # Remove this line for formal evaluation
)

run_task(task_cfg=task_cfg)

4.6 KiB Raw Blame History