Model chooser  /  Classify, tag or extract data  /  One-off

To tag or clean one list, paste it into Claude Sonnet 5.

A one-off cleanup of a few hundred rows does not need an API. Sonnet 5 is available to everyone in the Claude app, reads the CSV you upload, and is fast enough to iterate on the labels with you.

How to use it ↓ Try another task

Why Claude Sonnet 5

The reasons it wins for this job.

  • Anthropic lists Sonnet 5 as available to anyone on Claude.ai, so there is no plan to buy for a single job.
  • Classification is a task where the top models agree almost all the time. Speed of iteration, trying a label set and fixing it, matters more than the last point of accuracy.
  • Its 1M-token context holds a large CSV in one go, so labels stay consistent from the first row to the last.
  • It can write the result back as a CSV with a new column, ready to import into your CRM.

How to use it

4 steps to a first result.

  1. 1Define the labels in writingList every allowed value with a one-line definition and an example. Undefined labels drift.
  2. 2Run 20 rows firstCheck them by hand. Fix the definitions where it disagreed with you, then run the rest.
  3. 3Ask for a confidence columnHave it mark rows it was unsure about, and review only those.
  4. 4Export and spot-checkDownload the CSV and check another random 20 before importing.

Prompt for Claude

Start from this.

Edit the parts in capitals, then run it.

The attached CSV has one lead per row. Add three columns:

industry: one of SaaS, Fintech, Healthcare, Ecommerce, Agency, Other
company_size: one of 1-10, 11-50, 51-200, 201-1000, 1000+ (use the employees column; blank if unknown)
confidence: high or low

Definitions:
- SaaS: sells software on subscription
- ADD THE REST

Use only what is in the row. Never guess a company size. Do the first 20 rows and show me a table. After I confirm, do all rows and give me the full CSV.
Nothing is sent anywhere. It copies to your clipboard.

Alternatives that also work

If Claude Sonnet 5 is not an option.

GPT-6 AstraOfficial page →

OpenAI · closed · cost: high

Pick it when the list needs web lookups per row, such as finding each company's industry from its site.

Gemini 3.8 FlashOfficial page →

Google · closed · cost: low

Pick it when the data is inside a Google Sheet you already have open.

Gemma 4 26B A4BHugging Face →

Google · open weights · cost: free weights; you pay for the hardware · 9.6M HF downloads/mo

Pick it when the list contains personal data that cannot leave your laptop. It runs locally at 4-bit on a 24GB GPU.

Watch out for

Doing this again and again? To classify or extract at volume, use GPT-5.6 Luna. Tagging every inbound lead or ticket is a high-volume, low-difficulty job where cost per call decides the model. OpenAI built Luna for exactly that, it is in the lowest cost tier of any current flagship family, and strict JSON schemas keep the output machine-readable.

Sources

Checked . Models change monthly; we re-check this page when they do.

Picking the model is the easy part.

Wiring it into a workflow that runs every week, with evals, fallbacks and a cost you can predict, is the work. Fifteen minutes, no deck.

Book fifteen minutes →