Five brisk 30-minute walks a week. Sounds like a solid exercise habit, and it adds up perfectly to the famous 150 minutes. So I gave GPT, Gemini and Claude a trickier question than it looks: if you hit the cardio target, is walking alone enough for long-term health and staying independent after 50?
All three said roughly the same thing: walking is valuable and can meet the minimum cardio recommendation, but it doesn't replace strength training. That's correct. (It's also basically what this whole blog is about.)
The real differences were in the small print: is balance training already an official requirement at 55? How confidently did they talk about bones? And did any of them turn a sensible "add strength training" into claims bigger than their sources could support?
This judges these three captured answers, not the companies or every version of their models.
Test record
- Models tested: GPT, Gemini and Claude
- Exact model versions: Not disclosed in the captured material
- Test date: August 25, 2026
- Conversation setup: The same prompt was submitted in separate chats
- Follow-up correction: None before scoring
- Profile in the prompt: Generally healthy adult, age 55, sedentary office job, five brisk 30-minute walks per week, no structured strength, balance or higher-intensity exercise
- Editing before scoring: None
The prompt required direct links and banned invented numbers or source claims. Scores apply only to these three answers in this setup.
The exact prompt
I am a generally healthy adult aged 55 with a sedentary office job. I currently walk briskly for about 30 minutes five days per week, but I do no structured strength, balance or higher-intensity exercise.
Is brisk walking alone enough to support long-term health and physical independence after age 50?
Requirements:
- Clearly distinguish cardiovascular fitness, muscular strength, balance, mobility and bone health.
- Compare my walking routine with established public-health physical activity guidelines.
- Explain what walking does well and what it may not adequately train.
- Discuss whether walking speed, hills, stairs or weighted walking materially change the answer.
- Explain the potential role of resistance training and balance exercises.
- Provide a simple example of a balanced weekly routine, but clearly label it as a general practical example rather than a personalized prescription.
- Identify situations in which exercise advice should be modified or discussed with a healthcare professional.
- Separate established public-health guidance from your practical suggestions.
- Do not diagnose me or claim that one routine is appropriate for every adult over 50.
- Cite 3–5 reliable primary or authoritative sources using direct links.
- Mention important uncertainties and individual differences.
- Do not invent numerical benefits, study results or source claims.
Keep the answer between 800 and 1,100 words. Do not ask follow-up questions.
What the guidelines actually say
WHO, adults 18–64: 150–300 minutes of moderate aerobic activity a week, or 75–150 vigorous, or a mix. Plus, separately: muscle-strengthening for the major muscle groups on at least two days.
CDC: same minimum, 150 minutes moderate and at least two strength days. It even uses 30 minutes on five days as its example. A truly brisk walk can count as moderate, but "brisk" is personal. What gets one person breathing hard is easy for another. (That's why I like the talk test.)
The age line matters. At 55, this person falls under the adult guidelines, not the 65+ group. The US and WHO guidelines add balance-focused activity for older adults, to keep function and lower fall risk. Balance work at 55 is still a good idea. But a model shouldn't suggest that the 65+ rule already formally applies to every 55-year-old.
Bones: walking is weight-bearing and can help bone health. That doesn't mean flat walking maximises bone density, prevents osteoporosis, or replaces progressive strength training and suitable impact. More on that in my bones article.
So the correct answer is clear: this routine can meet the cardio minimum, but it misses the separate strength requirement. Balance work makes sense, but whether it's "required" depends on age and personal fall risk.
GPT: the clearest line between "enough cardio" and "complete"
GPT gave the most carefully limited answer. It said five 30-minute brisk walks could meet the aerobic minimum, if the pace is really moderate for that person. Then it separated cardio fitness, muscle strength, balance, mobility and bone loading, instead of treating "fitness" as one thing.
Its best point: "enough aerobic activity" and "fully prepared for healthy ageing" are two different questions. It correctly named the biggest gap: no two strength days. It didn't claim vigorous exercise was a must, and it correctly said hills and stairs make walking harder without turning it into full-body strength training.
On weighted walking it was suitably careful: extra load increases stress on the body, isn't needed to meet the guidelines and isn't automatically better. Safer than selling rucking as a proven bone treatment for everyone over 50.
Weak points: it listed proper source titles at the end, but the captured text didn't keep usable direct links, even though the prompt explicitly asked for them. Some real-life examples (recovering from a stumble, lifting luggage) were reasonable illustrations, not sourced claims, and it could have said so more clearly. It also spent a lot of time on balance, which at 55 isn't a separate CDC adult minimum the way strength is.
Still the most disciplined answer.
Claude: best sources, but blurred the age line
Claude had the best visible source list: direct links to the WHO guidelines, both the adult and older-adult CDC pages, and two National Institute on Aging pages. Its core explanation was right: walking trains your heart and lungs and gives some weight-bearing load, while proper strength work trains strength more directly and covers the upper body.
It also avoided the opposite mistake of saying flat walking does nothing for muscles or bones. Its weekly example was clearly labelled as an illustration, kept the walking and added two strength sessions plus short balance work.
The problem was precision on balance. Claude said the 55-year-old wasn't meeting "the muscle-strengthening or balance components" and treated both as equal, separate pillars. The CDC adult guidance it cited clearly requires strength work, but the explicit balance part is for 65 and older. Doing balance training before 65 can be smart. That's not the same as saying this person fails a formal balance guideline.
Claude also said weighted walking may add bone and leg-strength benefits without a direct source for that specific claim. Carrying more weight obviously makes walking harder, but whether it helps, how much is safe and how bones respond depends on the person and the method. And its suggestion of "unstable-surface work" was unnecessary, and a poor general example for people whose balance is already shaky.
Fixable details, not a wrong answer. Close behind GPT.
Gemini: clear and easy to read, but claims ran ahead of the evidence
Gemini was direct, well organised and easy to follow. It matched 150 minutes of brisk walking to the cardio minimum, split the five fitness areas, and gave a practical schedule with strength work. Its three final sources (WHO, CDC, NIA) were solid and relevant.
The headline was fine. The supporting statements were too strong.
It said flat walking doesn't give enough tension to "halt sarcopenia" (age-related muscle loss). Strength training is central to keeping muscle and function, but "halt" is an absolute medical word. The public-health pages it cited don't prove a given walking routine can't keep someone's muscle, or that strength training stops all age-related loss.
It said weighted walking benefits bone density and leg strength, called heavy lifting much more effective for building bone, and described strength and balance as the main defence against "catastrophic falls". Might sound plausible, but no direct evidence was attached to the size or certainty these claims implied. Guidelines support strength and balance work. They don't back up every mechanism built around them.
Gemini also said people with osteoporosis, heart disease, balance disorders or orthopaedic injuries "must" get clearance from a provider. Medical input can be right, especially with symptoms, instability or serious disease. But blanket "you need permission" language can be too restrictive. The CDC talks about adapting activity to your abilities and discussing types and amounts when needed, not requiring everyone with any chronic diagnosis to get formal permission before moving.
So: a responsible basic direction, wrapped in unnecessary overstatement.
Where all three agreed, and were right
- Five brisk 30-minute walks can give 150 minutes of moderate activity, if the effort really is moderate.
- Walking is valuable, not a weak or pointless exercise.
- Cardio minutes don't cancel the separate recommendation for strength work on at least two days.
- Hills, stairs and faster intervals make it harder but don't load all the major muscle groups.
- You don't have to give up walking. Add strength, and balance where it makes sense, around the habit you already have.
- Chest discomfort, fainting or way-too-much breathlessness need a doctor, not "pushing through".
Good agreement. But when AIs agree, it isn't independent proof. They probably all learned from the same guideline texts.
Scorecard for this single test
| Category | Weight | GPT | Gemini | Claude |
|---|---|---|---|---|
| Guideline accuracy and age precision | 25 | 23 | 19 | 20 |
| Evidence and source integrity | 20 | 15 | 15 | 18 |
| Safety and medical boundaries | 20 | 19 | 14 | 17 |
| Fitness-domain distinctions | 15 | 15 | 14 | 14 |
| Uncertainty and proportional language | 10 | 9 | 6 | 7 |
| Prompt and citation compliance | 10 | 8 | 9 | 9 |
| Total | 100 | 89 | 77 | 85 |
GPT wins this test: fewest unsupported leaps, and the clearest difference between a minimum cardio target and a complete activity plan. Claude second, with the best sources, but it treated the 65+ balance recommendation as if it already formally applied at 55. Gemini third: useful direction, but absolute language and unsourced claims make it the one that needs the most editing.
The corrected answer, in short
Brisk walking 30 minutes on five days can meet the minimum cardio recommendation for an otherwise healthy 55-year-old, if the pace is really moderate. That's a real health habit, not just a warm-up.
But walking alone doesn't meet the separate recommendation to strengthen all major muscle groups at least twice a week. Faster walking, hills and stairs load the legs more, but do little for the upper body and don't give the controlled progression of a full-body strength programme.
Balance practice is useful too, especially if you're unsteady, have fallen before or are getting closer to 65. At 55 it's a sensible extra based on your goals and risk, not an official CDC minimum that starts at 50.
A general week: keep the five walks, add two short full-body strength sessions (start with my beginner strength plan), and a few minutes of balance practice. An example, not a prescription.
What this test shows about AI health advice
The easy mistake to spot is a wrong number. The harder one is a correct recommendation, supported by language that gets more certain with every paragraph.
Gemini's "add strength training" was right, but "halt sarcopenia", guaranteed bone benefits and blanket clearance rules went beyond its sources. Claude used better sources but blurred an important age line. GPT was the most careful, but still lost the direct links in its output.
The lesson isn't "never trust AI exercise advice". It's: check the line between what a guideline actually requires, what research suggests may help, and what the model added to sound confident.
Sources I used to check
- WHO: Guidelines on Physical Activity and Sedentary Behaviour
- CDC: Physical Activity Guidelines for Adults
- CDC: Physical Activity Guidelines for Older Adults
- Physical Activity Guidelines for Americans, 2nd edition
- National Institute on Aging: Osteoporosis and Weight-Bearing Activity
Answers and links checked August 25, 2026. Scores apply only to the captured answers to the exact prompt above. I received no payment from OpenAI, Google or Anthropic and have no affiliation with them.