The latest version of ChatGPT has a new feature called Thinking Mode. (Which makes me wonder: What was it doing before?)
Thinking Mode takes more time to analyze, reason, and provide detailed information and insights.
Contrast it with Fast Mode, which is standard for most users. It gives you a speedy, efficient answer without a lot of reasoning. This is perfectly adequate for normal, everyday questions.
Thinking Mode is only available in the latest version, ChatGPT-5. You can turn it on and off manually, or it may be activated automatically by the complexity of your question.
Pro tip: if you’re using the free version (like most of us) and you’d like to engage Thinking Mode, turn it on before your first prompt. It seems that free users only get one Thinking Mode answer per day, so access is limited.
Thinking Mode vs. Fast Mode for writers and editors
Fast Mode is best for:
- Quick revisions or rewrites that don’t require deep analysis
- Polishing drafts for grammar, flow, or readability
- Simplifying text without rethinking structure or strategy
- Making suggestions for updating facts (but you still have to fact-check!)
- Generating multiple short variations, summaries, or headlines quickly
Thinking Mode is best for:
- Deep, structured writing and editing that requires analysis, synthesis, or strategic judgment
- Revising for audience fit, clarity of argument, or narrative flow
- Integrating or verifying research and evidence (again, though, you must fact check)
- Substantive rewriting (like reorganizing sections or aligning with brand voice)
Real-world example: Thinking Mode
I tested both Fast Mode and Thinking Mode with a paragraph of health writing for a lay audience. Here’s what ChatGPT came up with:
Prompt:
This content was revised two years ago. I want to 1) make sure the facts are up to date and 2) revise to make it more plain language. Show me how both modes would approach this task and what the outputs would look like.
Lecanemab and donanemab are FDA-approved immunotherapy drugs for the treatment of early Alzheimer’s. These drugs target the protein beta-amyloid to help reduce amyloid plaques, one of the hallmark brain changes in Alzheimer’s. Clinical studies to determine the effectiveness of lecanemab and donanemab were conducted in people with early-stage Alzheimer’s or mild cognitive impairment due to the disease. These studies showed that the drugs slowed the rate of cognitive decline among some study participants over the course of 18 months and reduced the levels of amyloid in the brain. Currently, insurance may only cover these medications in specific situations. Medicare Part B covers part of the cost of these medications for patients who meet certain medical criteria.
Fast Mode output:
Fast Mode Revised Paragraph:
lecanemab (brand name Leqembi®) and donanemab (brand name Kisunla™) are FDA-approved antibodies for people with early Alzheimer’s disease (that is, mild cognitive impairment due to Alzheimer’s or mild dementia due to Alzheimer’s). They work by targeting the protein beta-amyloid, which forms plaques in the brain that are a hallmark of Alzheimer’s disease. Clinical studies in people with early-stage Alzheimer’s showed that these drugs reduced the amount of amyloid in the brain and slowed cognitive decline in some participants. Under current Medicare policy, these medications may be covered for eligible patients when certain medical criteria are met and when providers participate in required data registries.
Why this is “fast”:
- I updated the FDA approval to “FDA-approved … for people with early Alzheimer’s disease” (fact-checked: lecanemab has full/traditional approval for early Alzheimer’s) Alzheimer’s Association, Biogen Investors, Alzheimer’s Disease International
- I added mention of Medicare policy with “when providers participate in required data registries” (fact: Medicare/National Coverage Determination requires registry participation) Alzheimer’s Association, Centers for Medicare & Medicaid Services, Managed Healthcare Executive
- Language is simpler and shorter.
- Doesn’t dive into “how much slower” or list detailed eligibility criteria.
My take:
- The good:
- The tone, level of detail, and word count are very similar to the original.
- ChatGPT gave me a breakdown of the changes it made, including links to outside resources for fact checking. The references it picked were reasonable and provided multiple sources of evidence to back up the statements.
- The not-so-good:
- The first sentence starts with a lower-case letter. It’s true that the generic names of these drugs aren’t capitalized, but it does need to be capitalized when it’s the first word in a sentence. That’s a pretty basic error.I don’t think “antibodies” is the most helpful way to describe the drugs. First, that term doesn’t identify them as drugs. Second, most folks aren’t going to know what antibodies are or what they do, so the term isn’t useful without additional explanation.I don’t like how it simplified the results of the clinical trial. It removed the 18-month time period, which I think is important—it would be misleading to imply that the improvement lasted forever.
- What happened to the information about medical insurance coverage?
Thinking Mode output:
Approach:
- Step 1: Review and fact-check each statement.
- Step 2: Consider audience (likely patients/families, or general public) and what plain-language goals appear.
- Step 3: Identify potential places to clarify or expand: e.g., what “early Alzheimer’s” means; what “slowed cognitive decline” means in practical terms; what coverage requirements involve.
- Step 4: Structure rewrite so that it moves smoothly from what the drugs are → how they work → what the evidence shows → what people should understand about coverage and eligibility.
- Step 5: Use plain language, define terms, avoid jargon or give simple definitions, add context about how much slowing was demonstrated or what “eligible” means.
Detailed Analysis:
- “Lecanemab and donanemab are FDA-approved immunotherapy drugs for the treatment of early Alzheimer’s.”
Fact-check: Lecanemab (Leqembi) has traditional FDA approval for early Alzheimer’s (mild cognitive impairment or mild dementia stage) in people with confirmed beta-amyloid pathology. Alzheimer’s Association, JAMA Network Donanemab (Kisunla) also has FDA approval for early Alzheimer’s disease. Lilly Investor Relations, Alzheimer’s Association So this is correct but could be clearer: what “early” means; that amyloid confirmation is required; that they are monoclonal antibodies.- Mechanism statement: “These drugs target the protein beta-amyloid to help reduce amyloid plaques, one of the hallmark brain changes in Alzheimer’s.”
This is accurate: they are anti-amyloid (beta-amyloid) antibodies. PMC, Alzheimer’s Association We might simplify “beta-amyloid” or explain plaques.- Evidence statement: “Clinical studies … showed that the drugs slowed the rate of cognitive decline among some study participants over the course of 18 months and reduced the levels of amyloid in the brain.”
Fact: Yes, the Phase III trial for lecanemab (Clarity AD) in ~1,795 participants showed statistical slowing of cognitive decline over 18 months. JAMA Network, American Academy of Neurology For donanemab, approval is based on early symptomatic Alzheimer’s, though we should mention that “some” means not everyone and effect size modest; also the amyloid reduction is shown. We might add a phrase like “by about 27% in some analyses” for context.- Coverage statement: “Currently, insurance may only cover these medications in specific situations. Medicare Part B covers part of the cost of these medications for patients who meet certain medical criteria.”
Fact: Medicare’s National Coverage Decision (NCD) requires that drugs for Alzheimer’s amyloid-directed monoclonal antibodies have traditional approval and coverage comes with “Coverage with Evidence Development” (CED) which includes patient registry participation. Managed Healthcare Executive So yes: Medicare Part B covers when criteria are met and provider participation in registry required. Also private insurance coverage may vary. We should clarify what “specific situations” means.Thinking‐Mode Revised Paragraph:
Two medications — lecanemab (brand Leqembi®) and donanemab (brand Kisunla™) — are now approved by the U.S. Food and Drug Administration (FDA) for people in the early stage of Alzheimer’s disease. Early stage means either mild cognitive impairment or mild dementia caused by Alzheimer’s, and a brain scan or test must show abnormal levels of a protein called beta-amyloid. Alzheimer’s Association, Lilly Investor Relations, American Academy of Neurology
These treatments are “monoclonal antibodies,” meaning they help the immune system remove beta-amyloid from the brain. Beta-amyloid plaques are sticky clumps of protein that build up in the brains of people with Alzheimer’s and are linked to damage of brain cells. PMC, Alzheimer’s Association
In clinical studies, some people who took lecanemab or donanemab saw less build-up of beta-amyloid in their brains and a slower drop in memory and thinking skills over about 18 months. For example, one study found a roughly 25–30 % slower decline in some participants compared with those who did not take the drug. JAMA Network, American Academy of Neurology
These drugs are not cures — they don’t stop Alzheimer’s or reverse it; rather they slow its progress in people who start them early and meet the medical eligibility criteria.
Regarding cost and insurance: For people with the standard Medicare program (Part B), coverage is available if both the doctor and patient agree to follow certain monitoring and data-reporting rules in a national registry. Alzheimer’s Association, Centers for Medicare & Medicaid Services Private insurance may have different rules. It’s important to check with your health plan and health-care team about whether you’re eligible and what costs might apply.
My take:
- The good:
- I like how ChatGPT described the details of its approach. The steps seemed reasonable and logical.
- The detailed analysis and reference links would make fact checking pretty straightforward. But I wouldn’t trust the information without checking it myself, even if ChatGPT says it’s doing a more careful job with Thinking Mode.
- Some of the plain language explanations are pretty nice. In some cases, it left in some more complex terminology without explaining it (e.g., cognitive impairment, eligibility criteria)
- I thought the call to action at the end was a helpful addition.
- The not-so-good:
- OY VEY, the “paragraph” it produced is actually FIVE paragraphs, more than doubling the word count of the original. If this were part of a larger article with a word limit, that would NOT work.
- If I were going to refine this output, I’d use a subsequent prompt to limit the scope and word count of the rewrite.This revision also added tons of details that might not match the scope of the rest of the piece.
- It missed a few subtle but important details, like the fact that early-stage Alzheimer’s disease and mild cognitive impairment are not the same thing.
For what it’s worth, Copyleaks.com (AI detector) identified no AI-generated content in either output.
Conclusions:
A casual user of ChatGPT probably doesn’t need to worry about Thinking Mode. Let the model do its thing, and you’ll get a serviceable answer most of the time. In fact, Thinking Mode might give you WAY MORE information than you want or need, actually making it less useful.
And its detailed steps and strategizing might lull you into thinking that it’s doing all the fact checking for you; let me assure you that you must still do your due diligence there.
But it’s fun to play with. For paid users of ChatGPT, it would make sense to make it part of your regular writing and editing process. Just keep in mind what it’s helpful for and what it’s not, and don’t let it fool you into thinking it’s smarter than it is!