---
title: "Compare GPT, Claude and Gemini side by side"
slug: "compare-gpt-claude-gemini-answers"
type: "apify-task-example"
url: "https://automationbyexperts.com/apify/examples/compare-gpt-claude-gemini-answers"
actor: "Run GPT, Claude, DeepSeek... Prompts in Bulk (No API Key)"
actor_url: "https://automationbyexperts.com/apify/bulk-llm-runner"
apify_url: "https://apify.com/fayoussef/bulk-llm-runner/examples/compare-gpt-claude-gemini-answers?fpr=youssef&fp_sid=abe_ex_compare-gpt-claude-gemini-answers"
price: "$0.002"
price_unit: "per result"
---

# Compare GPT, Claude and Gemini side by side

The same prompts sent to GPT-5, Claude Sonnet 5, Gemini 3.8 Flash and Perplexity Sonar, returned as one row per model per prompt with the answer, the cost and the token count.

**Run it:** [Compare GPT, Claude and Gemini side by side on Apify](https://apify.com/fayoussef/bulk-llm-runner/examples/compare-gpt-claude-gemini-answers?fpr=youssef&fp_sid=abe_ex_compare-gpt-claude-gemini-answers)  
**Actor:** [Run GPT, Claude, DeepSeek... Prompts in Bulk (No API Key)](https://automationbyexperts.com/apify/bulk-llm-runner)  
**Pricing:** $0.002 per result

## What this task is for

The three sample prompts are deliberately different jobs: an explanation for a non-technical reader, a tone rewrite of a support reply, and a reasoning question about churn and growth. Models that look equal on a benchmark table separate quickly across that mix.

Cost and token count sit next to each answer, so the comparison covers what a model charges as well as what it writes. Replace the prompts with the ones your product actually sends before choosing what goes into production.

## The input this task runs with

```json
{
  "prompts": [
    "Explain what a vector database is to a non technical product manager, in under 120 words.",
    "Rewrite this support reply to be warmer without adding promises: 'Your refund was denied because the return window closed.'",
    "A SaaS company has 4% monthly churn and 8% monthly new customer growth. Explain in plain language whether it is growing, and what breaks first."
  ],
  "compare_models": [
    "openai/gpt-5",
    "anthropic/claude-sonnet-5",
    "google/gemini-3.8-flash",
    "perplexity/sonar"
  ],
  "system_prompt": "You are a helpful, accurate assistant. Answer directly and concisely.",
  "response_format": "text"
}
```

## What every run returns

- `Prompt`
- `Model used`
- `Response text`
- `Structured JSON columns`
- `Web search sources`
- `Source row (Excel/CSV/Sheet)`
- `Status`
- `Timestamp`

## How to run it

1. Open the task on Apify. New accounts get $5 of free credit.
2. Check the pre-filled input above and change anything you need.
3. Click Run. Pagination, proxies and anti-bot handling are automatic.
4. Export CSV, JSON or Excel, or save it as your own task and schedule it.

## Other Run GPT, Claude, DeepSeek... Prompts in Bulk (No API Key) tasks

- [Write Amazon listing copy for every SKU in a spreadsheet](https://automationbyexperts.com/apify/examples/amazon-listing-copy-from-product-spreadsheet): A full Amazon listing for every SKU in your spreadsheet: title, five bullets, description and backend search terms, each in its own column, written only from the facts in your sheet.
- [Run a ChatGPT prompt for every row of a Google Sheet](https://automationbyexperts.com/apify/examples/run-chatgpt-prompts-from-google-sheet): One ChatGPT prompt per row, run for a whole Google Sheet or Excel file. Paste the link, get the file back with each answer in a new column. No API key of your own required.
- [Enrich a company list with founded year, HQ, CEO and size](https://automationbyexperts.com/apify/examples/enrich-company-list-with-ai): Upload a sheet of company names and websites, ask your question in plain words, and get the answers back as new columns: founded year, headquarters, CEO and employee count.
- [Check which tests each lab offers from an Excel list](https://automationbyexperts.com/apify/examples/check-lab-tests-from-excel-list): Upload a list of laboratories and a checklist of tests, and get yes, no or unknown for every lab and test, each answer with an evidence sentence and a source link.
- [Classify and tag customer reviews at scale](https://automationbyexperts.com/apify/examples/classify-customer-reviews): Every review through the same rubric: sentiment, a topic tag, an urgency flag and a one-line summary, each as its own column rather than a paragraph of commentary.
- [Write SEO product descriptions in bulk](https://automationbyexperts.com/apify/examples/seo-product-descriptions-in-bulk): One line per product in, a structured description out: a benefit-led paragraph, feature bullets and a meta description, as JSON columns ready for a product feed.

---

Part of [AutomationByExperts](https://automationbyexperts.com) - see [all example tasks](https://automationbyexperts.com/apify/examples) or [request a new one](https://automationbyexperts.com/suggest).