A beginner does not need a universal model ranking. You need a tool that is available in your region and account, fits the software you already use, and is approved for the data you plan to give it. ChatGPT, Claude, Gemini, and Microsoft Copilot are reasonable products to test, but their features, limits, and prices change frequently.
Below is the version of this comparison that does not list 90 features. Read it and pick one. You can always change your mind.
The short version
| If you… | Use |
|---|---|
| …live in Gmail, Google Docs, and Drive | Test Gemini’s eligible Workspace connections |
| …live in Outlook, Word, and Excel | Test the Copilot entitlement in your Microsoft account |
| …mainly draft and analyse long text | Test Claude alongside one alternative |
| …want a general-purpose assistant with configurable GPTs and several media tools | Test ChatGPT |
| …need current facts with inspectable sources | Use a search-enabled tool and open the cited sources |
Use the table to choose candidates, not a winner. Features and account permissions still need verification.
ChatGPT (OpenAI)
ChatGPT is a general-purpose assistant with web search, file analysis, image generation, voice features, projects, and custom GPTs. Availability and limits depend on the plan; check the current ChatGPT plan comparison rather than relying on a fixed feature list in an article.
Useful fit: you want one general-purpose interface and may use custom GPTs, file analysis, images, or voice.
What to test: source faithfulness, editing effort, enabled connectors, sharing defaults, and whether your organization approves the account for the intended data.
Cost: a free plan exists. OpenAI listed Plus at USD 20 per month when this article was checked on 4 August 2026; local taxes, currency conversion, features, and limits can differ. Verify before paying.
Keep ChatGPT on the shortlist if the specific models, tools, data controls, and Custom GPT permissions you need are present in your account.
Claude (Anthropic)
Claude is a general-purpose assistant with a text- and document-oriented workflow. Projects can hold project instructions and reference material. Treat claims about “best writing” or “best reasoning” as preferences, not facts: test the kind of document you actually produce.
Useful fit: you work mainly with drafting, critique, long documents, or code and prefer its interface and output in a side-by-side test. Check Anthropic’s Claude plan documentation for current availability and limits.
What to test: whether the current plan supports the media, tools, knowledge, memory, and sharing behavior required for your workflow. Do not infer parity with ChatGPT or Gemini from similar feature names.
Cost: free and paid tiers exist, but price and usage limits vary by region and can change. Verify the checkout price and cancellation terms in your account.
Keep Claude on the shortlist if a side-by-side test on your own low-risk writing or documents shows lower correction effort and the account meets your data requirements. Do not use this criterion to delegate legal interpretation.
Gemini (Google)
Gemini can connect to Google services such as Gmail, Docs, Drive, Calendar, Keep, and Tasks, subject to account type, region, settings, and administrator controls. Google’s own documentation warns that connected-app answers can be outdated or wrong, so open the listed source email or file before acting (Google’s connected-app guidance).
What to test: eligible Workspace connections, source links back to Gmail or Drive, NotebookLM, supported media, context limits, and administrator controls.
Product-boundary risk: Gemini Apps, AI Overviews, NotebookLM, API models, and Workspace features are different surfaces with different grounding, activity, and data behavior. Confirm which one you are using.
Cost: free and paid Google AI plans exist. Do not treat an AI upgrade as “free” because you already buy storage: compare the actual incremental price, included storage, account eligibility, and cancellation effect on the current Google AI plans page.
Keep Gemini on the shortlist if your eligible Google account exposes the required connections and they pass your source and privacy checks.
Copilot (Microsoft)
Microsoft uses the Copilot name for several consumer and work products. Microsoft 365 Copilot features can appear in Word, Excel, Outlook, PowerPoint, and Teams, but entitlements, administrator settings, and supported actions vary. Check the Microsoft 365 Copilot overview for the account you will actually use.
What to test: eligible Excel, Outlook, Word, PowerPoint, or Teams features; tenant policy; citations back to source material; and the exact data-protection terms for the organization account.
Product-boundary risk: consumer Copilot and organization-managed Microsoft 365 Copilot are not interchangeable. Compare the exact entitlement rather than the shared name.
Cost: consumer and business offers change. Check your Microsoft account or employer entitlement rather than assuming Copilot is included in an existing Microsoft 365 licence.
Keep Copilot on the shortlist if the employer has approved the intended use and the specific Microsoft 365 features reduce measured correction or transfer work.
A brief word on Grok and the rest
Grok (xAI) is closely integrated with X. That can make it useful for examining discussion on X, but social posts are not verified evidence. Product personality and guardrail comparisons are subjective and change too quickly to use as a safety or quality criterion.
Perplexity is a search-oriented answer product and one candidate when you want links alongside an answer. Its citations can still be irrelevant, incomplete, or misrepresented; open the primary sources before relying on a claim.
Meta AI, DeepSeek, Mistral Le Chat, and others may also fit. Their availability, hosting, privacy terms, and capabilities differ. Include them only if they pass the same task, data, and account checks; brand popularity is not a quality test.
A test you can do this afternoon
Pick the same task — a real one, ideally something you would actually do today — and run it through two of the four options. Compare the outputs side by side. Notice:
- Which one’s tone feels closer to how you would have written it?
- Which one asked a clarifying question (a good sign) versus just generating something?
- Which one’s interface you found less annoying?
After the test, record which output required fewer factual corrections and less editing. Use that evidence, not the model’s self-description, to choose.
The differences matter less than the habit
Online model rankings rarely represent your account, data boundary, source needs, or task. Even an everyday drafting or summarization result can require substantial correction.
Choose from a small, low-risk comparison and record correction effort. Pay only when the live plan solves an observed constraint, and reassess when the products, policy, or work change.
It is also reasonable to stay on a free plan or use ordinary search. Pick one low-risk task, test two tools if needed, and review the choice when your work or the products change.
Sources checked on 4 August 2026
OpenAI’s ChatGPT pricing and plan comparison and GPT availability documentation; Anthropic’s Claude plan documentation; Google’s documentation for connecting Workspace apps to Gemini and the Gemini Apps Privacy Hub; and Microsoft 365 Copilot’s product page.
