Anthropic's own site lists five Claude plans, seven products and four model families. Knowing how they fit together helps you avoid paying for the wrong tier.
Claude is not one thing, and Anthropic's own site shows it. The company lists five plans, seven products and four model families, and the names overlap. The easiest way through is this: a product is a place you go, a plan is what you are allowed to do there, and a model is what answers.
Anthropic's FAQ names them: Free, Pro, Max, Team and Enterprise. The pricing page splits them into two groups. Free, Pro and Max are individual plans. Free is for everyone, Pro is for everyday work, and Max is for people who work with Claude all day. Team and Enterprise are priced per seat, meaning per person, and are aimed at groups.
The main menu lists Claude, Claude Code, Claude Cowork, Claude Duet and Claude Tag. Under a heading called Specialized sit Claude Security and Claude Science. Separately, nine capabilities, including Artifacts, Connectors, Skills and Claude in Chrome, are things Claude does or places it appears inside other software. They are not destinations.
The menu is still changing. A banner on the pricing page reads "Claude Cowork is now just Claude. Rolling out to Pro and Max, with more plans to follow," while the navigation still lists both names as of 23 October 2026.
The families are Mythos, which includes Fable, plus Opus, Sonnet and Haiku. The versions listed on 23 October 2026 are Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5. Sonnet and Haiku are included on every plan, including Free. Opus is marked No on Free and Yes from Pro upward. Fable is also marked No on Free.
Anthropic states prices plainly for developers who use Claude in their own software. These are per million tokens, the units the company bills in:
The subscription side is far less clear. Anthropic does not say how many tokens a $20 Pro plan buys. The published half is precise, and the unpublished half is the subscription most people actually pay for.
Check the product, the plan and the model separately before you pay: Opus needs Pro or higher, and Anthropic does not say how much usage Pro buys.
Visa tested 22 AI models on realistic shopping tasks. They passed just over half, and often bought the wrong thing anyway.
Read more: Visa tested 22 AI shoppers, and most still make costly mistakesDARPA wants AI to hunt for design flaws in stock, bond, energy and cloud markets, so enemies cannot use them to cause economic damage that reaches households.
Read more: Pentagon wants AI to find weak spots in stock and energy marketsTuring Award winner Richard Sutton says more computing power alone will not make AI truly smart. Big AI companies disagree, which matters for anyone betting on AI.
Read more: AI pioneer says piling on computing power won't make machines smart
