DOABLE · QUICK GUIDE
Which AI brain do I pick?
Model menus change constantly. This decision rule does not: match the amount of thinking to the importance of the task.
THE REAL-LIFE PROBLEM
What this guide helps you do
Model menus use changing labels such as Fast, Thinking, Pro, Sonnet, or Opus without clearly explaining which amount of reasoning your task needs.
WHEN THIS HELPS
Use it when a task is complex, consequential, or unfamiliar enough that you are unsure whether the default setting is sufficient.
WHEN TO SKIP IT
Skip it for a quick everyday question, a low-stakes draft, or a simple lookup you will verify anyway.
Let AI help you decide
ONE MAIN PROMPT
CONTEXT My task is: [describe the task and why the answer matters]. ROLE Act as a practical guide to choosing the right amount of AI reasoning for this task. INTERVIEW If the stakes, deadline, or complexity are unclear, ask me up to 3 short questions, one at a time, before recommending a setting. TASK Tell me whether this task needs a fast response, balanced reasoning, or deeper reasoning, and explain why in plain language. If a different model or setting would help, tell me what label to look for before answering. Then list the tradeoffs, open questions, and facts I should double-check myself.
You will get back: a short recommendation on whether the task needs speed or careful reasoning, plus a plain reason why.
The five-second decision rule
Ask one question: does this task need a quick answer, or more careful reasoning?
Quick answer: use the fast option
Everyday questions, a first draft, a definition, brainstorming, or something you will double-check anyway. Look for labels such as Instant, Fast, Flash, or Haiku.
More careful reasoning: use the deeper option
Use more reasoning for multi-step problems, complex comparisons, or anything that will shape an important decision. Look for labels such as Thinking, Pro, Opus, or extended reasoning.
Not sure? Start with the recommended setting
Some tools can choose or increase reasoning automatically. Availability depends on the product and plan. If you are unsure, keep the recommended setting or use the prompt below.
Fast vs. balanced vs. deep
| Speed | What it is good for | Example task |
|---|---|---|
| Fast | Everyday questions, quick drafts, brainstorming, and simple lookups. | Give me five dinner ideas using chicken and rice. |
| Balanced | Most day-to-day writing, planning, organizing, and explaining. | Help me reply to this email from my landlord. |
| Deep | Multi-step problems, complex comparisons, and consequential work you will verify. | Compare these insurance plans and list what I should confirm before choosing. |
The deeper option is not always better. It is usually slower, and it can overthink a simple question.
What the names mean, without the technical part
The companies behind popular AI tools use different labels for a similar idea: faster everyday responses and more deliberate reasoning for harder tasks. Exact choices depend on your plan and change often.
- ChatGPT: Current ChatGPT guidance describes Instant for everyday work, with eligible paid plans able to increase reasoning automatically. Manual reasoning levels and Pro availability depend on the plan.
- Claude: Anthropic describes Haiku as its fastest model, Sonnet as the balance of speed and intelligence, and Opus for complex agentic work. Its current lineup also includes Fable as a highest-capability model, so the choices you see may differ.
- Gemini: Google currently labels Gemini 3 Flash choices as Fast or Thinking, while Pro is for more complex work. Higher limits and some options depend on the subscription.
You do not need version numbers. Remember the practical distinction: use the fast label for a quick answer and the reasoning or pro label when the task needs more careful thought.
BEFORE YOU ACT
What to verify yourself
What AI can do
- Recommend fast versus deep mode for a task.
- Increase reasoning automatically when supported.
- Explain why a task may need more depth.
What you decide
- Decide whether the task is quick-answer or right-answer territory.
- Override the automatic choice when needed.
- Decide whether the deeper answer was worth the wait.
Sources and details
- OpenAI Help Center, GPT-5.6 in ChatGPT, for current Instant, reasoning, and plan guidance.
- Anthropic, Models overview, for the current Claude lineup and intended use cases.
- Google, Gemini Apps release updates, for current Fast, Thinking, Flash, and Pro labels.
Model names and capabilities change frequently. The speed-versus-depth decision rule is designed to remain useful even when labels change.
Last checked: August 3, 2026
You do not need to understand every model name. You need to know whether your task needs speed or depth.