On 3 March 2025, Google launched its new Data Science Agent tool, powered by Gemini 2.0 artificial intelligence, on the Colab platform. The new AI agent automates data analysis processes and generates complete Jupyter notebooks based on simple natural language instructions. Integrated into Colab, the tool is primarily designed for data scientists and AI developers, but it also assists in identifying API anomalies, analysing customer data, and writing SQL code.
The new Data Science Agent represents the integration of a previously standalone project into Colab, significantly reducing the time required for data analysis. As Kathy Korevec, Product Director at Google Labs, stated in an interview, we’re only scratching the surface of what people will be able to do. A researcher from Berkeley National Laboratory found that data processing time dropped from one week to just five minutes using the tool. The Data Science Agent ranked 4th on the Hugging Face DABStep test, outperforming several other AI tools, including ReAct agents built on GPT 4.0, Deepseek, Claude 3.5 Haiku, and Llama 3.3 70B.
The Data Science Agent is available for free in certain countries and languages, though the free version of Colab offers limited computational capacity. Paid Colab plans are offered at four different tiers: Colab Pro ($9.99/month) with 100 compute units, Colab Pro+ ($49.99/month) with 500 compute units, Colab Enterprise with Google Cloud integration, and a Pay-as-you-go option providing 100 compute units for $9.99 or 500 compute units for $49.99. The tool currently supports CSV, JSON, or .txt files under 1GB and can process approximately 120,000 tokens per request, automating tasks such as importing libraries, loading data, and generating basic code.
Sources:
1.
Carl Franzen
2.
Jane Fine
3.
Jagmeet Singh
