Text and reasoning
General quality for writing, structuring ideas, reviewing tone and explaining decisions.
Practical data
AI prices, plans and benchmarks for better decisions
Prices can be published if they come from public sources, but must include date and official link. This page was reviewed on 27/08/2026 and uses indicative amounts: always confirm the final price, currency, VAT and limits before paying.
Quick reading
Start by: Free trial with ChatGPT, Claude, Gemini, Perplexity or Copilot if you already have it included. For payment: Search for a plan of around 20 USD/EUR per month only if the task is repeated. For business: Look at privacy, administration, users and if data is used for training.
Benchmarks help filter options, but do not replace a test with your emails, documents, budgets or real code.
Visual view
Quick charts by real use
Editorial 0 to 10 scale based on public benchmarks, product limits and practical fit. Use it to filter what to test, not as an absolute verdict.
Code and agents
Indicative performance for editing projects, solving bugs and working on real files.
Long documents
Reading, synthesis, traceability and context handling for reports, contracts or quotes.
Search with sources
Ability to answer with visible links, recent information and quick context checks.
Generalist Virtual Assistants Plans
| Tool | Free Plan | First estimated payment. | Team/company | When to test | Link |
|---|---|---|---|---|---|
| ChatGPT | Yes | Plus/Pro: confirm on the official website. OpenAI publishes Business from €21/user/month and Premium from €105/user/month. | Business and Enterprise, with administration, connectors, and corporate privacy. | General option for writing, reasoning, analyzing files and testing various tasks. | Prices ChatGPT |
| Claude | Yes | Pro/Max: confirm on the official website according to country and limits. | Team and Enterprise. | Long writing, careful review, extensive documents and delicate tone. | Claude pricing |
| Gemini | Yes | Google AI Plus/Pro/Ultra: the price varies by country; includes storage and wide access to Gemini. | Gemini Enterprise for work accounts. | Google Workspace, Gmail, Drive, long documents and Google ecosystem. | Planes Google AI |
| Microsoft Copilot | Depends on account and product | COPilot Pro or Microsoft 365: confirm in Microsoft. | Microsoft 365 Copilot for business. | Microsoft Word, Microsoft Excel, Microsoft Outlook, Microsoft Teams and already added organizations in Microsoft 365. | Microsoft Copilot |
| Perplexity | Yes | Pro/Max: confirm on the official website. | Enterprise Pro. | Search with visible sources, news, and quick comparisons. | Prices Perplexity |
Programming tools
| Tool | Free | First estimated payment. | Clear usage | Link |
|---|---|---|---|---|
| Cursor | Free hobby | $20/month; Teams $40/user/month. | Modify real projects with agent, border models and editor context. | Prices Cursor |
| GitHub Copilot | Free | Pro $10/user/month; Pro+ $39; Max $100. | Programming within editors and GitHub, with chat, agent mode, and reviews. | Plans Copilot |
| Codex | According to plan OpenAI | Included in plans compatible with ChatGPT/OpenAI, or via API depending on usage. | Code agent for reading files, editing, executing tests, and explaining changes. | OpenAI Codex |
| Claude Code | According to plan Anthropic | Confirm in Claude/Anthropic. | Code-assisted work by Claude on projects and terminal. | Claude Code |
Automation and processes
| Tool | Initial plan | Useful data | When it fits | Link |
|---|---|---|---|---|
| Zapier | Free | Free with 100 tasks/month; payment plans depend on level and task volume. | Connect apps without programming: forms, email, CRM, sheets and notices. | Zapier pricing |
| Make | Free | Variable price per operations and plan; confirm on official website. | Automation with more technical control than a simple template. | Prices Make |
| n8n | Cloud or self-hosted | Starter, Pro, Business and Enterprise plans; the website highlights projects, concurrent executions, history and credits of AI. | Technical teams wanting more control or hosting their own flows. | n8n pricing |
Trustworthy Benchmarks You Should Look At
Artificial Analysis
Compares intelligence, speed, latency, task cost and context cost. Useful for spotting trends, not for choosing a subscription without testing it.
View leaderboardLM Arena
Ranks models with human preferences in conversational comparisons. Serves for general quality, but does not always reflect documents, company or costs.
View leaderboardSWE-bench
Programming benchmark with real software incidents. Useful if the goal is code, agents, and bug resolution.
View SWE-benchAider LLM Leaderboards
Code editing tests with estimated cost. In the reviewed version showed GPT-5 at the top of its main table, with Gemini, o3 and Claude also among strong options according to configuration.
View Aider benchmarkHow to use these data without making mistakes
Choose two candidate tools, open official links, confirm price and limits, and test the same real case on both. For a small business, the winner is the one that reduces reviewable time: fewer repeated emails, fewer reading errors, clearer budgets or code that can be tested.