jupyter-agent/jupyter-agent-qwen3-4b-instruct
Jupyter Agent Qwen3-4B Instruct is a 4 billion parameter causal language model, fine-tuned from Qwen3-4B-Instruct-2507 by Jupyter Agent, specifically for data science agentic tasks within Jupyter notebook environments. This model excels at executing Python code, analyzing datasets, and providing clear reasoning for realistic data analysis problems. It features robust code execution capabilities with popular data science libraries and achieves state-of-the-art performance for small models on the DABStep benchmark, demonstrating significant improvement over its base model.
Loading preview...
Jupyter Agent Qwen3-4B Instruct Overview
Jupyter Agent Qwen3-4B Instruct is a specialized 4 billion parameter language model, fine-tuned from Qwen3-4B-Instruct-2507, designed to function as a data science agent within Jupyter notebooks. Developed by Jupyter Agent, this model is optimized for executing Python code, performing data analysis, and generating clear, reasoned explanations for complex data science problems. It leverages a fine-tuning process on a unique dataset of synthetic notebooks and Kaggle workflows to enhance its agentic capabilities.
Key Capabilities
- Jupyter-native agent: Seamlessly integrates into notebook environments for interactive data science.
- Advanced Code Execution: Capable of executing Python code using libraries like pandas, numpy, and matplotlib for data manipulation and visualization.
- Clear Reasoning: Provides intermediate computations and explanations, making its analytical process transparent.
- Dataset-Grounded Analysis: Trained on realistic Kaggle notebook workflows, enabling effective analysis of diverse datasets.
- Tool Calling: Supports structured code execution and final answer generation through defined tools.
Performance Highlights
On the DABStep benchmark for data science tasks, Jupyter Agent Qwen3-4B Instruct demonstrates significant improvements:
- Achieves 70.8% on Easy Tasks, up from 44.0% for the base model.
- Achieves 3.4% on Hard Tasks, up from 2.1% for the base model.
- Represents a 36.3% improvement over the base model and 22.2% improvement over scaffolding alone, positioning it as a state-of-the-art small model for data analysis.
Good for
- Automating data analysis workflows in Jupyter notebooks.
- Generating Python code for data manipulation, visualization, and statistical analysis.
- Users requiring clear, step-by-step reasoning for data science tasks.
- Developing intelligent agents for data exploration and insight generation.
- Applications where a small, efficient model with strong data science capabilities is needed.