Claude.ai Basics

Claude is an AI assistant built by Anthropic, accessible primarily through a web and mobile chat interface at claude.ai. Understanding a few basics helps you use it more effectively.

Models and tiers. Anthropic offers several models that balance speed and capability differently — historically named things like Haiku, Sonnet, and Opus, roughly in order of increasing depth. A lighter model responds faster and suits quick everyday tasks, while a more capable model handles research, complex writing, or multi-step reasoning better. Anthropic periodically introduces new tiers and versions, so the exact lineup available to you depends on your plan.

Plans. Claude.ai offers a free tier with usage limits, along with paid plans (commonly Pro and higher tiers) that raise usage limits and unlock access to more capable models, longer conversations, and additional features.

Core features. Beyond plain chat, Claude.ai includes several features worth knowing:

• Projects let you group related conversations together with shared context, files, and custom instructions — useful for ongoing work rather than one-off questions.
• Artifacts let Claude generate a standalone piece of content — code, a document, a diagram — that renders in its own panel next to the conversation, so you can iterate on it without losing the surrounding chat.
• Memory (where available) lets Claude retain useful context about you or your projects across separate conversations, instead of starting from zero every time.
• File uploads let you share documents, images, or spreadsheets directly in a conversation for Claude to read and work with.

Context window. Every conversation has a limit on how much text — yours and Claude's combined — it can hold at once, called a context window. Very long conversations can eventually hit this limit, which is one reason Projects and Memory exist: to carry relevant context forward without re-pasting it every time.

Knowing these basics — the model tiers, what each plan unlocks, and what Projects, Artifacts, and Memory actually do — is the foundation for using Claude well, and it's exactly what this topic's questions test.