Five free plans, five different personalities. Picking one isn't about which AI is “best,” it's about which gap it's least likely to leave you with.
ChatGPT: the generalist
Free ChatGPT is the safest default if you want one tool for everything: writing, quick research, images, voice conversations. Its voice mode is the most natural-sounding of the three, and it has the broadest feature set on a free plan. The trade-off is that the free tier gives you a lighter model most of the time, with limited access to its flagship reasoning model before it throttles back (Towards AI; T-Minus AI).
Claude: the writer
Claude consistently scores best on writing quality and instruction-following, the kind of output that reads like a person wrote it rather than an AI ticking boxes (Towards AI; Artificial Corner). It's also the strongest coder of the three on real-world benchmarks. The catch on the free plan is volume: Claude's free tier runs out of messages faster than ChatGPT or Gemini, so it suits shorter, higher-value sessions better than all-day back-and-forth.
Gemini: the researcher
Gemini's edge on a free plan is real-time information. It pulls in current web results more consistently than the other two, and it produces the strongest images of the three (Towards AI). Its free tier is also the most generous on usage volume and context window size, useful if you're feeding it long documents. Where it lags is writing polish: output leans functional over natural.
Copilot and Grok: the other two
Microsoft Copilot rounds out the five platforms CORDON works with. Its free-tier standout is unlimited web search, no visible daily cap where the other four eventually throttle, and it's tightly integrated with Word, Outlook, and Teams if you're already living in that ecosystem.
Grok has grown fast: from roughly 35 million monthly active users in December 2025 to around 117 million by March 2026 (Fatjoe), already ahead of Claude by user count and the third most-used assistant in the US behind ChatGPT and Gemini. Its edge is real-time access to X, useful when a question depends on what's happening right now, less useful for most everyday writing and planning tasks.
At a glance
| Model | Best for | Free-tier trade-off |
|---|---|---|
| ChatGPT | All-round daily use, voice, images | Flagship model access is limited before it steps down |
| Claude | Writing quality, coding, following detailed instructions | Runs out of messages fastest of the three |
| Gemini | Current information, long documents, image generation | Writing quality is the weakest of the three |
| Copilot | Unlimited web search, Microsoft ecosystem | Less useful without Word, Outlook or Teams already in your workflow |
| Grok | Real-time X/social context | Weaker fit for everyday writing and planning |
Free-tier limits shift most months, sometimes within the same month, so treat the table above as directional rather than exact. Whichever changed last, the shape stays the same: one model per weakness, not one clear winner.
The model isn't the bottleneck
Pick any of the five above and the same six questions still apply, the model changes, your situation doesn't. A prompt built with CORDON works unchanged in ChatGPT, Claude, Gemini, Copilot or Grok, so switching models to chase a strength from the table above never means starting over. Why AI can't help without context. covers why that input matters more than the model choice itself.