Machines vs free weights: does it matter?

Machines are not cheating. In studies that compared them head to head, machines and free weights built about the same strength and muscle.^[1][2] The one steady difference: people got best on the kind of equipment they trained with.^[1][2] So use what you like, and train the exact lift you want to get better at.

The answer isn't obvious because the gym floor has strong views. Free weights are "real" lifting. Machines are for beginners, or for people who are scared of the squat rack. The studies are calmer than that.

The short version is in gym basics. This article goes further. It covers the "stabiliser muscle" argument, why your numbers change when you switch equipment, safety and cables.

What getting this wrong costs you

If you think machines don't count, you might skip the training you'd actually do. Or you might load a barbell lift you haven't learned yet, because it feels like the serious option.

The opposite mistake costs something too. Say you move from a machine to a barbell and the number drops. It's easy to think you've lost strength. Usually you haven't. You've started a new skill, and the research below shows why the number changes.

What the research says

A few terms first. Free weights are weights you hold and move freely: barbells, dumbbells and kettlebells. Machines guide the weight along a fixed path, like a leg press or a Smith machine (a barbell fixed on rails). A meta-analysis pools the results of many studies on the same question into one estimate. Hypertrophy means muscle growth: the muscle fibres get bigger. Your one-rep max (1RM) is the heaviest weight you can lift once.

Head to head: no clear difference

For strength, that's moderate evidence. For muscle size it's limited evidence. Only 5 of the 13 studies in the 2023 meta-analysis measured size, and mostly in indirect ways such as tape measurements and skinfolds.^[1]

The two meta-analyses asked the same question about two years apart. So we count them as one line of evidence, not two.

A position stand is an official summary of the evidence written by a panel of experts. The American College of Sports Medicine (ACSM) published one in 2026. It found that equipment type didn't change strength gains, and said there wasn't enough data to judge muscle size.^[3] It based both on the 2023 meta-analysis alone, so it isn't a separate line of support.^[3]

Three trials up close

Three trials are worth a closer look. Two of them measured muscle directly with ultrasound, a scan that shows how thick a muscle is. They found the same pattern.

• Spain. 38 men who already trained were allocated to barbells or machines for 8 weeks. They did the squat, bench press, shoulder press and a row lying face down on a bench. The programs differed only in the equipment. Strength and muscle size rose about the same in both groups.^[4]
• Canada. 46 young adults, 26 of them women, were randomised (put into groups by chance) to free weights or machines for 8 weeks. Their biceps and front thighs grew about the same. Strength on the free-weight squat, free-weight bench press and Smith-machine squat rose 11–19% in both groups.^[5]
• Belgium. 36 men new to lifting trained for 10 weeks. One group used machines, one used free weights, and one switched from machines to free weights after 5 weeks. All three gained about the same size, strength and score on a movement test. Switching halfway cost nothing.^[6] Limited evidence on its own: a small trial that measured size with a tape.

Two of these trials are already in the 2023 meta-analysis. The Spanish one is new and points the same way. Strength stays at moderate evidence: each trial is short, and none is large.

The Canadian trial also checked hormones. In men, free weights raised free testosterone more right after a session. Muscle and strength still rose about the same in both groups.^[5] Limited evidence: one trial. A bigger hormone spike after a workout isn't a reason to pick free weights.

You get best at the kind you train

Specificity means your body gets better at the exact task you practise.

Both meta-analyses found it. People who trained with free weights gained more strength on free-weight tests. People who trained on machines gained more on machine tests. On a neutral test, the gains were about the same.^[2] In the 2023 meta-analysis the gaps were small, and the machine half was only a hint, not a finding.^[1] The 2022 one found bigger gaps, of a moderate size.^[2]

The detail is shakier than the headline. In the 2023 meta-analysis, the free-weight half leaned heavily on one large study of 304 men. Without it, that gap was no longer clear.^[1]

Newer trials point the same way. In a 2025 study, 37 sport-science students (14 women) trained for 8 weeks. They did the squat and bench press, with free weights or on a Smith machine. Strength rose more when the test matched how they had trained. Muscle grew about the same in both groups.^[7] The groups were matched for sex and strength, and strength was estimated from bar speed rather than a 1RM test.^[7] In the Spanish trial, both groups got better on both kinds of test, "but especially in the one they trained".^[4]

That's moderate evidence for specificity. Rep ranges and load covers a related idea.

Do stabiliser muscles tip the balance?

The classic case for free weights goes like this. A barbell wobbles, so smaller "stabiliser" muscles have to work to steady it. More muscles working should mean more growth.

Part of that is true in the moment. EMG (electromyography) measures the electrical activity in a muscle with sensors on the skin. In one study, average muscle activity was 43% higher in a free-weight squat than in a Smith-machine squat. Three leg muscles were more active. The lower back and the abs were not.^[8] That study had just 6 people, who each did one set of 8 reps.

More activity in one set hasn't been shown to predict more growth over months. A 2022 review argued that EMG readings are not a validated way to predict muscle growth.^[9] Expert consensus: this is the authors' view, not a tested finding. The long-term trials above measured growth itself, and found about the same with either kind.^[4][5]

Making a lift wobblier didn't help either. A 2015 meta-analysis looked at 22 studies of training on unstable surfaces, like balance boards. Fifteen of them compared it with the same training on stable ground. In teenagers and young adults, the wobble added little to strength, power or balance.^[10] Limited evidence: the studies were of low quality, the authors call their findings preliminary, and none compared the two in older adults.^[10]

Are machines safer?

Nobody has studied this well yet. We found no trial that compared injury rates between machines and free weights. In the Spanish trial, joint pain and stiffness went down in both groups, by about the same amount.^[4] Limited evidence: 38 trained men over 8 weeks is far too few to count injuries.

So we can't say machines are safer. We can't say free weights are as safe, either. Learning the lifts covers what injury records do show, and how to learn a free-weight lift with a light load.

What about cables?

A cable machine has a handle on a cable attached to a weight stack. Unlike most machines, it doesn't fix the path you move along. Nobody has studied this well yet. The 2023 meta-analysis left cables out, so it can't tell us how cables compare.^[1]

How to apply it

This is coaching practice, not a tested finding. The steps follow from the studies above, but no trial tested them as a plan.

1. Use the equipment you'll actually use. For general strength and muscle, the studies found no clear winner.^[1][2] Pick by what you like and what your gym has.
2. Train the lift you want to get better at. Want a stronger barbell bench press? Bench with a barbell. Strength gains lean toward the equipment you train on.^[2][7]
3. Start on machines if a barbell lift is new to you. Men who switched from machines to free weights halfway through ended level with the rest.^[6] Start the free-weight lift light.
4. Mix both if you like. Nothing in these trials says you have to choose one.
5. Skip the balance gadgets. Training on unstable surfaces added little over stable ground.^[10]
6. Log a machine lift and its free-weight version as two exercises. Their numbers don't compare.

Which exercises to pick, and how far each rep should move, is in choosing exercises and range of motion.

A worked example

Say a 55 kg lifter trains the machine chest press for 8 weeks. (These numbers are illustrative, not from a study.) She ends at 35 kg × 10, 10, 9. Then she switches to the dumbbell bench press. Her first session reads 10 kg dumbbells × 10, 10, 8.

That looks small next to 35 kg. It isn't lost muscle. A chest press and a dumbbell press are different skills, and strength is partly specific to the movement.^[2][7] In the Belgian trial, the men who switched ended level with the others after 10 weeks.^[6]

So she logs the dumbbell press as its own exercise. Next week she aims to beat it by one rep: 10 kg × 10, 10, 9.

Common mistakes

Calling machines cheating. Gym culture treats the barbell as the real test. In the studies, machines built about the same strength and muscle.^[1][2] Use the machine without apology.

Thinking you have to move off machines once you've trained a while. Machines can look like training wheels. But training experience didn't change the results in the 2023 meta-analysis, and the Spanish trial used men who already trained.^[1][4] Switch when you want to get better at a free-weight lift, not because of a rule.

Thinking machines don't build "real" strength. The idea is that a fixed path can't carry over to life or sport. A second paper from the same Spanish team used 34 trained men on the same kind of program. It found no difference between groups in jumping or other athletic tests.^[11] The Belgian movement test came out about the same too.^[6] If a sport or task matters to you, practise that task as well.

Comparing numbers across equipment. 35 kg on a chest press and 10 kg dumbbells aren't the same test. Compare each exercise only with itself.

How Gymbers fits

Each version of a lift needs its own log. The Gymbers app logs the weight and reps for every set, whether the exercise uses a machine or a barbell. Each exercise has its own history, charts and records, so a machine chest press and a dumbbell bench press each show their own progress. Exercises come with a video, which helps when a free-weight lift is new. The app marks new PRs (personal records) in the workout summary.