Neither pricing page settles which is better, but they settle what you pay: ChatGPT Plus and Claude Pro are both $20 a month in the US. Pick ChatGPT if you want image creation on every tier, custom GPTs and a cheaper entry tier; pick Claude if you want Claude Code and Research in the same plan and a usage model built around five-hour sessions. Both have a free tier, and both let you opt out of model training. We compared the vendors’ published plans, not the answers each writes.
Price and what the paid tier adds
| Attribute | ChatGPT | Claude |
|---|---|---|
| Free | $0: text chats, limited uploads, image creation, voice, deep research, memory | $0: Sonnet and Haiku models, web search, files and code execution, memory, artifacts, up to 5 projects |
| Entry paid | Go, $8 a month: more messages, uploads, image generation, voice and memory | none below Pro |
| Standard paid | Plus, $20 a month, monthly billing only: advanced reasoning models, extended limits, extended Codex; the pricing page also lists projects, scheduled tasks and custom GPTs (OpenAI’s Go help article says Go has those too) | Pro, $20 a month or $17 a month billed $200 yearly: adds Opus, Claude Code, Research, Claude Design, Slides and Docs, Claude in Chrome, Microsoft 365, unlimited projects |
| Heavy use | Pro, $100, $200 or $500 a month | Max, $100 (5x Pro usage) or $200 (20x) a month |
| Image generation | Listed on every tier | Not listed on the pricing page |
| Context window | Reasoning models 256K tokens on Go and Plus, 400K on Pro; 27K to 128K for the fast models | “Up to 1M, varies by model” on every plan |
Sources: OpenAI’s help centre and Go announcement for the dollar prices; the Netherlands pricing page (euros: Go 8, Plus 23, Pro from 103, a regional price) for the tier contents and context windows; Claude’s pricing page and Max help article. Business, Enterprise and Team plans are not covered.
How limits work
Claude’s page says every plan has a usage limit that resets every five hours, paid plans add weekly limits, and web, desktop, mobile and Claude Code all draw on one pool. There is no fixed message count. ChatGPT’s tier table says text chats on Free are unlimited “as long as there is no abuse”, with other features carrying limits it does not number. Neither page publishes a message cap, so neither can be compared on numbers.
Pick by what you need
- Pick ChatGPT if image creation, custom GPTs or a cheaper entry tier (Go at $8) matter: Claude’s pricing page lists no image generation and no custom-GPT store, and has no tier below $20. Both list scheduled tasks.
- Pick Claude if you want Claude Code, Research and Microsoft 365 and Chrome integrations inside a $20 plan, or you read very long documents and want the longer context Anthropic lists on every plan.
- Stay free on both if you ask occasional questions: each free tier includes web search and file handling.
- Look elsewhere for cited research (Perplexity) or Google-app integration (Gemini).
What this page cannot tell you
Writing quality, accuracy, instruction following and coding results are not on a pricing page, and we have not assessed them. Earlier versions of this comparison ranked the two on those points; the sources were not there, so the verdicts were removed. Read the full plan pages in ChatGPT review and Claude review.
Method
Researched from OpenAI’s help centre, Go announcement and Netherlands pricing page, and Anthropic’s pricing page and help centre, all read 29 Sep 2026. Nothing here comes from using either product.