Foundations · Chapter 6 of 11
Standard, reasoning & agentic
Not all AI models are the same. Understand the difference between reasoning, creative, and specialist models — and when to use which.
Most tools now offer a few modes. You don't need the technical detail — just to know which to reach for, and when.
Fast and direct
These respond immediately. Use them for most tasks like drafting, summarising, brainstorming, or reformatting. They're great for Q&A against your own documents.
Slow and deliberate
These models "think" before they respond. This can take up to a minute. They're designed to break complex problems into logical steps and use more compute power.
Reasoning modes are slower and cost more. Start with a standard model and step up only when the problem genuinely needs the extra depth.
A third kind: models that take action
The newest shift is agentic tools. Rather than just answering, they can use other software, browse, work across your files and run multi-step tasks on your behalf — checking their own work as they go. You give a goal; they do the legwork.
- General assistants: ChatGPT, Claude, Gemini, Microsoft Copilot
- Desktop/file assistants such as Cowork that act across your apps
- Coding agents: OpenAI Codex, Claude Code, Cursor, Gemini CLI
- Same underlying models — the difference is what they're allowed to do
Chat tools tell you how. Agentic tools go and do it — which is exactly where the next chapter picks up.