1 of 8

5-MINUTE LEARNING SESSION

Make Co-Pilot Work for You:

Choosing the Right Model

Auto, Quick Response, Think Deeper, and the GPT models. Which one to pick, and when to let Copilot decide.

GenAI Literacy Series

2 of 8

Where to Find the Model Picker

Open Copilot Chat. The model picker is the dropdown labelled "Auto" in the top-right of the chat window. Click it to see your options.

Response modes

Tell Copilot how much time to spend.

Auto

Decides how long to think

Quick Response

Answers right away

Think Deeper

Thinks longer for better answers

Specific GPT models

Pick the engine yourself. Found under "More" in the dropdown.

GPT 5.3 Quick Response

Fast everyday model. Newer than 5.2.

GPT 5.4 Think Deeper

Deepest reasoning. Best first drafts.

GPT 5.2 Quick Response

Legacy fast model.

GPT 5.2 Think Deeper

Legacy reasoning model.

If you don't see the picker: Microsoft is still rolling this out. The exact options also depend on your subscription and region.

3 of 8

How Auto Picks for You

Auto is the default. It uses something called dynamic model routing.

Your prompt

Whatever you typed

Auto router

Reads your prompt, decides complexity

If simple → Quick Response (fast model)

If complex → Think Deeper (reasoning model)

Signals that route to Quick

  • Short, factual questions
  • Quick rewrites or tone adjustments
  • Simple summaries of one document
  • "Translate this to French"

Signals that route to Think Deeper

  • Multi-step problems or trade-off analyses
  • Long context or multiple sources to synthesize
  • Strategy, planning, or recommendations
  • Open-ended questions with no obvious answer

Bottom line: Auto handles maybe 80% of everyday work well. Override it when speed or depth genuinely matters.

4 of 8

Quick Response

Answers right away. Same model Auto sends simple prompts to. Skip the routing decision when you already know what you need.

Drafting & rewrites

Example prompts:

"Polish this email so it sounds less stiff."

"Rewrite this paragraph in plain language."

Why Quick:

You know what good looks like. Don't need extra reasoning, need flow.

Translation & quick edits

Example prompts:

"Translate this paragraph to French."

"Fix the grammar in this sentence."

Why Quick:

Routine work where speed matters more than depth.

Quick Q&A

Example prompts:

"What does GCDocs stand for?"

"Who's the Director of [team]?"

Why Quick:

Factual answers you'd otherwise Google. Faster to ask Copilot.

5 of 8

Think Deeper

Spends up to ten seconds reasoning before answering. Better at multi-step problems, trade-offs, and longer context. Worth the wait when accuracy matters.

Trade-off analysis

Example prompts:

"Compare three vendors on cost, risk, and timeline. Recommend one."

"Should we phase this rollout or do it all at once? Pros and cons."

Why Think Deeper:

Multiple variables, no single right answer. Reasoning helps.

Strategy & planning

Example prompts:

"Draft a six-month rollout plan with milestones and risk flags."

"Help me think through the stakeholders for this initiative."

Why Think Deeper:

Output needs to be structured, defensible, and complete.

Synthesis across sources

Example prompts:

"Summarize these four reports and pull out the common themes."

"Read this 30-page document and tell me what's missing."

Why Think Deeper:

Long context. Pattern recognition across multiple documents.

6 of 8

When to Pick a Specific GPT Model

Most of the time, you don't. But here's what each one is good at if you want control.

GPT 5.3 Quick Response

Fast, current

The most current fast model. Better instruction-following and writing than 5.2. Pick this when you want speed but want the newest improvements over the older fast model.

Best for: Everyday drafting, summaries, quick rewrites where you want quality without a wait

GPT 5.4 Think Deeper

Deepest reasoning

The most capable reasoning model in the picker. Stronger first drafts on complex tasks. Better at long, multi-step work. Worth the extra wait when the output really matters.

Best for: Strategy documents, complex analysis, anything where you'd usually iterate three or four times to get it right

GPT 5.2 Quick Response

Legacy fast

Older fast model. Still works fine. Mostly there for consistency if you started a project on 5.2 and want the output to match.

Best for: Continuity with older work. Otherwise, prefer 5.3.

GPT 5.2 Think Deeper

Legacy reasoning

Older reasoning model. Solid on complex tasks but superseded by 5.4 Think Deeper for new work.

Best for: Continuity with older work. Otherwise, prefer 5.4 Think Deeper.

Quick rule: If you can't tell which model to pick, you don't need to. Auto handles it. Override only when you have a reason.

7 of 8

Decision Cheat Sheet

Match the task to the mode. If you're not sure, leave it on Auto.

Task you have in front of you

What to pick

Why

Polish an email before sending

Quick Response

Fast, no reasoning needed

Translate a paragraph

Quick Response

Routine work, speed wins

Quick factual question

Quick Response

Don't wait for reasoning

Compare three options with trade-offs

Think Deeper

Multi-variable analysis

Plan a project rollout

Think Deeper

Multi-step, structured output

Summarize a long document

Think Deeper

Long context, synthesis

Mixed bag of small tasks

Auto

Let the router decide

Not sure how complex this is

Auto

Default works most of the time

You can change the mode mid-conversation. If your first answer feels shallow, switch to Think Deeper and re-ask.

8 of 8

ACTION LIST

Before you close this tab

One session is enough to feel the difference.

1

Open Copilot. Click the dropdown in the top-right. See what's actually in your version.

2

Run a simple prompt on Quick Response. Run the same prompt on Think Deeper. Compare the answers.

3

Try one complex task on Think Deeper that you'd normally re-prompt three times. See if you get there in one shot.

4

Leave it on Auto for the rest of the week. Override only when you have a clear reason.

Picking the right model isn't about knowing GPT versions. It's about knowing whether you need a fast answer or a careful one.