reasoning benchmark
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.
Updated Aug 7, 2026
Published models1
Registry coverage1
MetricScore
EvidenceB
DS-Arena-Code leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | DE | 63.1% | 100.0% | 1 | C |
DS-Arena-Code highlights
The top published results on this benchmark's own scale.
What is DS-Arena-Code?
Definition and scoring fields from the benchmark registry.
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.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- DS-Arena-Code
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- ds-arena-code|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about DS-Arena-Code.
Which model scores highest on DS-Arena-Code?
DeepSeek-V2.5 is currently ranked first with 63.1%.
What does DS-Arena-Code measure?
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.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
1 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.