# DS-STAR: A state-of-the-art versatile data science agent

DevFeed: [DS-STAR: A state-of-the-art versatile data science agent](<https://devfeed.tech/articles/ds-star-a-state-of-the-art-versatile-data-science-agent-6761.md>)

Original publisher: [Read original article](<https://research.google/blog/ds-star-a-state-of-the-art-versatile-data-science-agent/>)

Published: 2025-11-06T17:50:36Z

Content type: article

Language: en

Sources: [The latest research from Google](<https://devfeed.tech/sources/the-latest-research-from-google.md>)

Topics: [Data Science](<https://devfeed.tech/topics/data-science.md>), [Data Mining & Modeling](<https://devfeed.tech/topics/data-mining-modeling.md>), [data-processing](<https://devfeed.tech/topics/data-processing.md>), [LLM evaluation / benchmarking](<https://devfeed.tech/topics/llm-evaluation-benchmarking.md>), [Benchmark](<https://devfeed.tech/topics/benchmark.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [code](<https://devfeed.tech/tags/code.md>), [data-mining-modeling](<https://devfeed.tech/tags/data-mining-modeling.md>), [data-processing](<https://devfeed.tech/tags/data-processing.md>), [data-science](<https://devfeed.tech/tags/data-science.md>), [google](<https://devfeed.tech/tags/google.md>), [json](<https://devfeed.tech/tags/json.md>), [large-language-models-llms](<https://devfeed.tech/tags/large-language-models-llms.md>), [machine-intelligence](<https://devfeed.tech/tags/machine-intelligence.md>), [markdown](<https://devfeed.tech/tags/markdown.md>), [natural-language-processing](<https://devfeed.tech/tags/natural-language-processing.md>), [statistics](<https://devfeed.tech/tags/statistics.md>), [verification](<https://devfeed.tech/tags/verification.md>), [visualization](<https://devfeed.tech/tags/visualization.md>)

## AI overview

DS-STAR is a data science agent from Google Cloud that automates tasks including statistical analysis, visualization, and data wrangling across varied data types. It uses file analysis, LLM-based verification, and iterative sequential planning to produce verifiable insights, achieving state-of-the-art results on the DABStep, KramaBench, and DA-Code benchmarks.

## Source excerpt

Data Mining & Modeling