DS-Arena-Code

Data Science Arena Code benchmark for evaluating LLMs on realistic data science code generation tasks. Tests capabilities in complex data processing, analysis, and programming across popular Python libraries used in data science workflows.

DeepSeek-V2.5 from DeepSeek currently leads the DS-Arena-Code leaderboard with a score of 0.631 across 1 evaluated AI models.

Paper
About this benchmark

What DS-Arena-Code measures

DS-Arena-Code is a text benchmark that evaluates large language models on reasoning tasks. LLM Stats tracks 1 model on this benchmark, with a maximum possible score of 1. Current average across reported models is 0.6, with the leader reaching 0.6.

Compare leaders on the best AI for reasoning leaderboards.

Publication

Paper
DSCodeBench: A Realistic Benchmark for Data Science Code Generation
Authors
Shuyin Ouyang, Dong Huang, Jingwen Guo, Zeyu Sun, and 2 others
Published

Abstract

We introduce DSCodeBench, a new benchmark designed to evaluate large language models (LLMs) on complicated and realistic data science code generation tasks. DSCodeBench consists of 1,000 carefully constructed problems sourced from realistic problems from GitHub across ten widely used Python data science libraries. DSCodeBench offers a more challenging and representative testbed, more complex code solutions, more comprehensive data science libraries, clearer and better structured problem descriptions, and stronger test suites. To construct the DSCodeBench, we develop a robust pipeline that combines task scope selection, code construction, test case generation, and problem description synthesis. The process is paired with rigorous manual editing to ensure alignment and enhance the reliability of the evaluation. Experimental result shows that DSCodeBench exhibits robust scaling behavior, where larger models systematically outperform smaller ones, validating its ability to distinguish model capabilities. The best LLM we test, GPT-4o, has a pass@1 of 0.392, indicating that LLMs still have a large room to improve for realistic data science code generation tasks. We believe DSCodeBench will serve as a rigorous and trustworthy foundation for advancing LLM-based data science programming.

DeepSeekDeepSeek-V2.5 leads with 63.1%.

Progress Over Time

Interactive timeline showing model performance evolution on DS-Arena-Code

State-of-the-art frontier
Open
Proprietary

DS-Arena-Code Leaderboard

1 models
ContextCostLicense
1236B
Notice missing or incorrect data?

FAQ

Common questions about DS-Arena-Code.

What is the DS-Arena-Code benchmark?

Data Science Arena Code benchmark for evaluating LLMs on realistic data science code generation tasks. Tests capabilities in complex data processing, analysis, and programming across popular Python libraries used in data science workflows.

What is the DS-Arena-Code leaderboard?

The DS-Arena-Code leaderboard ranks 1 AI models based on their performance on this benchmark. Currently, DeepSeek-V2.5 by DeepSeek leads with a score of 0.631. The average score across all models is 0.631.

What is the highest DS-Arena-Code score?

The highest DS-Arena-Code score is 0.631, achieved by DeepSeek-V2.5 from DeepSeek.

How many models are evaluated on DS-Arena-Code?

1 models have been evaluated on the DS-Arena-Code benchmark, with 0 verified results and 1 self-reported results.

Where can I find the DS-Arena-Code paper?

The DS-Arena-Code paper is available at https://arxiv.org/abs/2505.15621. The paper details the methodology, dataset construction, and evaluation criteria.

What categories does DS-Arena-Code cover?

DS-Arena-Code is categorized under reasoning. The benchmark evaluates text models.

What is the best open-source model on DS-Arena-Code?

DeepSeek-V2.5 by DeepSeek is the top-ranked open-source model on DS-Arena-Code, with a score of 0.631 (rank #1).

How recent are the DS-Arena-Code leaderboard results?

The DS-Arena-Code leaderboard was last updated in June 2026 and currently includes 1 evaluated models.

More evaluations to explore

Related benchmarks in the same category

View all reasoning
GPQA

A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.

reasoning
224 models
MMLU-Pro

A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks. Features over 12,000 curated questions across 14 domains and causes a 16-33% accuracy drop compared to original MMLU.

reasoning
127 models
AIME 2025

All 30 problems from the 2025 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999. Used as an AI benchmark to evaluate large language models' ability to solve complex mathematical problems requiring multi-step logical deductions and structured symbolic reasoning.

reasoning
114 models
MMLU

Massive Multitask Language Understanding benchmark testing knowledge across 57 diverse subjects including STEM, humanities, social sciences, and professional domains

reasoning
100 models
SWE-Bench Verified

A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.

reasoning
100 models
Humanity's Last Exam

Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions

reasoningmultimodal
82 models