ChatGPT Code Interpreter: What It Became and How Data Analysis Works (2026)

ChatGPT Code Interpreter began as a plugin in 2023, but it is no longer something most users need to install. Its capabilities evolved into ChatGPT’s built-in data analysis tools. Today, ChatGPT can inspect uploaded spreadsheets and other supported files, run Python-based calculations, create tables and produce charts, depending on the model, plan and workspace settings.

If you have found an old guide telling you to install a “Code Interpreter plugin” in an editor, ignore those instructions. The current workflow is to upload a supported file to ChatGPT and describe the analysis you want. This guide explains the change and how to use data analysis responsibly.

What happened to ChatGPT Code Interpreter?

Code Interpreter was introduced as a ChatGPT plugin in 2023. It allowed ChatGPT to run code in a controlled environment and work with uploaded files. The feature later became part of ChatGPT’s native data-analysis experience, commonly referred to as Advanced Data Analysis. OpenAI’s current Help Centre describes the feature as “Data analysis with ChatGPT”.

The important distinction is that the old plugin installation steps are obsolete. You do not need to download a Code Interpreter extension for Visual Studio Code or another code editor to use ChatGPT’s file-analysis features. Availability is controlled by ChatGPT’s current interface, model, plan and workspace settings.

What can ChatGPT data analysis do?

• Inspect uploaded spreadsheets and summarise rows, columns, trends and unusual values.

• Clean, transform, filter or combine structured data.

• Calculate totals, averages, percentages and other statistics.

• Create tables and charts to make results easier to understand.

• Run Python-based calculations and statistical analysis for suitable tasks.

• Explain the method, assumptions and results in plain language.

How to analyse a spreadsheet with ChatGPT

1. Open a ChatGPT conversation where file uploads and data analysis are available.

2. Upload a supported spreadsheet, such as an XLSX or CSV file.

3. Describe the question you want answered. Name the relevant columns, date range, filters and calculations.

4. Ask ChatGPT to show the method and any assumptions it makes.

5. Check the result against the original data before using it in a report or decision.

For example, you could upload a monthly sales spreadsheet and ask: “Group revenue by month, calculate the percentage change from the previous month, identify the three largest product categories and create a line chart. Show the calculations and flag any missing values.” A specific request is easier to check than asking simply to “analyse this file”.

Which file types can it analyse?

OpenAI lists common data-analysis file types including Excel workbooks (.xlsx), CSV files, PDFs and text or data formats such as JSON, XML, YAML, TXT and Markdown. The exact formats and file limits can vary by model, plan, workspace settings and current upload allowance.

For spreadsheets, use descriptive column headings, one record per row and consistent data types. Avoid merged cells, unrelated tables on the same sheet and empty rows that split the data. If a PDF contains scanned pages or image-based tables, extraction may be less reliable than working from the original spreadsheet.

Does ChatGPT really run Python?

For some data-analysis tasks, ChatGPT writes and runs Python code in a stateful, sandboxed environment. This allows it to perform calculations and transformations rather than only describing code that a person could run elsewhere. The environment used for data analysis cannot make external web requests or API calls, so analysis that depends on live external data needs an uploaded dataset or an available connected source.

You should still review the code, outputs and assumptions before relying on a result. Ask ChatGPT to show the calculation, explain how it handled missing values, or repeat the analysis using a method you specify. A chart can look convincing even when the underlying data or assumptions are wrong.

What are the limitations?

• Large, complex or poorly structured files may not be analysed completely.

• Scanned PDFs and image-based tables can be difficult to extract accurately.

• The model may choose a method or chart type that does not match your intent.

• A calculation can be technically correct but answer the wrong business question if the data or assumptions are misunderstood.

• File limits and feature availability vary by account and workspace.

For important financial, operational, scientific or compliance decisions, reconcile key totals with the source file and have a suitably qualified person review the method. Do not treat a generated chart or narrative summary as independent verification.

Privacy and uploaded files

Only upload data you are permitted to share. Before using ChatGPT with customer, employee or commercially sensitive information, check your organisation’s policy and the privacy settings for your account. OpenAI explains that uploaded files may have separate retention from the conversation, especially when files are saved to Library or a Project. See the official Chat and file retention guidance.

Code Interpreter vs a traditional coding environment

ChatGPT data analysis is useful when you want to ask questions about a file in natural language and receive calculations, tables or charts. It is not a replacement for a full development environment when you need to build, test, deploy and maintain production software. For repeatable business reporting, keep a record of the source data, code, formulas and validation checks.

Frequently asked questions

Is Code Interpreter still a ChatGPT plugin?

No. Code Interpreter began as a plugin in 2023, but its capabilities evolved into built-in ChatGPT data analysis. The old plugin installation instructions are no longer the right way to access the feature.

Can ChatGPT analyse Excel files?

Yes. ChatGPT can analyse supported Excel workbooks such as XLSX files, as well as CSV files. Availability and limits depend on the account, model and workspace.

Can ChatGPT create charts from data?

Yes. ChatGPT can create tables and charts from suitable uploaded data. Check the data, grouping and calculation behind a chart before using it.

Can ChatGPT analyse a scanned PDF?

It may not reliably extract exact values from scanned PDFs or image-based tables. If exact numbers matter, use the original spreadsheet or a text-based PDF and verify extracted values.

Official source and further reading

OpenAI’s current guide, Data analysis with ChatGPT (updated 18 days ago), explains supported file types, analysis tasks, Python execution, charts and limitations. For retention, see Chat and file retention in ChatGPT.

Last reviewed: 7 November 2026. We will update this page if OpenAI changes the data-analysis experience or supported file formats.