What is the most accurate wearable for VO2 max?
Garmin ranks first for VO2 max estimation, followed by Fitbit and Google. The ranking compares wearable estimates with laboratory gas-analysis measurements and gives the most weight to independent exercise-based validation and direct same-protocol comparisons.
VO2 max accuracy by brand
VO2 max on a wearable is an estimate rather than a direct oxygen-consumption measurement. The ranking therefore prioritizes comparisons against laboratory indirect calorimetry or spirometry, with direct head-to-head studies carrying the most weight.
Rank | Brand | Verdict |
|---|---|---|
1 | Strongest independent evidence, including several exercise-based validations and a direct head-to-head win | |
2 | Independent replication across exercise-based studies with roughly 9 to 10% error | |
3 | Strong current Pixel Watch 3 result, but the evidence still rests on one funded validation | |
4 | Broad historical evidence, but higher error in direct and current validation studies | |
5 | Several recent independent validations, with consistently larger error than the leaders | |
NR | Promising manufacturer validation, but not yet independently replicated | |
NR | Cardio Capacity is available, but peer-reviewed criterion validation is not yet strong enough | |
NR | Insufficient current peer-reviewed VO2 max criterion-validation evidence | |
NR | Insufficient VO2 max criterion-validation evidence for a numbered rank | |
NR | Insufficient current connected evidence for a defensible numbered rank |
How the ranking was decided
Wearable VO2 max estimates are useful for tracking fitness, but they are not interchangeable with a laboratory test. The INTERLIVE systematic review and meta-analysis found that exercise-based algorithms generally performed better than resting estimates, while individual error could still be substantial even when average bias was small.
Garmin ranks first because the result is replicated across several independent studies and populations. Fenix 6 produced MAPE of about 7% in both the general-population validation and the athletic-population validation, while Forerunner 245 was around 7 to 8% overall in the runner study. Most importantly, the same-participant Garmin versus Polar comparison reported 7.3% MAPE for Garmin Forerunner 920XT versus 13.2% for Polar V800.
Fitbit ranks second on evidence depth rather than having the single lowest error figure. The larger Charge 2 exercise validation reported roughly 9% MAPE with small average bias, while a separate independent Charge 2 study reported 10.2% MAPE and strong agreement. Those are older devices, but the replication is valuable.
Google ranks third because Pixel Watch 3 produced one of the strongest current-model results. In the 2026 Pixel Watch 3 validation, MAPE was 8.65% and average bias was small, although the individual limits of agreement were wide. The study was Google-funded and is currently the main criterion-validation result supporting the brand, so Google stays behind Fitbit until the finding is independently replicated.
Polar ranks fourth. Its broader historical evidence is substantial, but the direct Garmin comparison found 13.2% MAPE for Polar V800, while the 2025 Vantage M2 validation reported about 13.7% MAPE for the resting Polar Fitness Test. That is useful for trend tracking, but weaker than the connected evidence above it.
Apple ranks fifth despite having some of the most current independent research. Apple Watch Series 7 showed 15.79% MAPE in one criterion validation, a broader 2025 Apple Watch study reported 13.31%, and Series 10 testing was around 13.2%. The consistency of the evidence raises confidence in the conclusion, but not in a higher rank.
The 2025 qualitative systematic review reinforces the broader pattern. Most included studies evaluated Garmin watches and Firstbeat-based estimates, and seven of 13 studies judged wearable VO2 max estimation valid or acceptable. The review also warned that accuracy is less dependable in elite endurance athletes, so a brand rank should not be read as a guarantee for every fitness level.
WHOOP remains unranked even though its internal laboratory validation reported 8.0% MAPE for the passive model and 7.1% with GPS augmentation across 248 maximal treadmill tests. That is promising, but manufacturer-run validation is supporting evidence rather than a substitute for independent replication. Oura, Suunto, Samsung and Huawei also remain unranked because the connected peer-reviewed criterion evidence is not strong enough for a numbered position.
Supporting studies
GarminGarmin Forerunner 245, Garmin Fenix 6, Garmin Forerunner 920XT4 studies
Garmin ranks first on the strongest connected evidence base. Several independent Fenix and Forerunner studies report exercise-based VO2 max errors below 10%, and a direct same-participant comparison put Garmin materially ahead of Polar.
- Exercise / athletes August 6, 2025Validity of V̇O2max estimates from the forerunner 245 smartwatch in highly vs. moderately trained endurance athletes
Forerunner 245 underestimated laboratory VO2max by 4.73 and 4.05 mL/kg/min across two outdoor runs. Overall MAPE was 7.9% and 7.2%, improving to 4.1-2.8% in moderately trained athletes but worsening to 10.4-9.4% in highly trained athletes.
- Laboratory January 1, 2025Validation of Aerobic Capacity (VO2max) and Pulse Oximetry in Wearable Technology
Garmin Fenix 6 VO2max estimates aligned closely with laboratory values. MAPE was 7.05% against 30-second averaged VO2max, with Lin concordance 0.73 and bias +0.75 mL/kg/min.
- Laboratory May 26, 2023Validation of Aerobic Capacity (VO2max) and Lactate Threshold in Wearable Technology for Athletic Populations
Garmin Fenix 6 VO2max estimate had MAPE 6.85% against the 1-minute laboratory VO2max average, Pearson correlation 0.81 and Lin concordance 0.70. Error was larger against shorter laboratory averaging windows.
- Laboratory August 22, 2019Validity of Wrist-Worn Activity Trackers for Estimating VO2max and Energy Expenditure
Garmin Forerunner 920XT had MAPE 7.3%, MAE 3.8 mL/kg/min and ICC 0.82. It significantly underestimated VO2max by about 2.1 mL/kg/min but was materially more accurate than Polar V800 in the same participants.
FitbitFitbit Charge 22 studies
Fitbit ranks second. Two independent Charge 2 studies reported roughly 9 to 10% mean absolute percentage error, giving Fitbit useful replication rather than relying on a single favorable result.
- Laboratory September 23, 2019Assessing the Ability of the Fitbit Charge 2 to Accurately Predict VO2max
Fitbit Charge 2 overestimated VO2max by about 2.92 mL/kg/min on average; MAPE was 10.2% and ICC 0.87.
- Exercise / athletes May 21, 2019Validity of Cardiorespiratory Fitness Measured With Fitbit Compared to VO2max
Fitbit Charge 2 showed small mean bias against laboratory VO2max and MAPE of 9.41% versus 15-second laboratory averaging and 9.14% versus 60-second averaging. Fitness-category agreement improved to 91.7% when grouped into three clinical bands.
GoogleGoogle Pixel Watch 31 study
Google ranks third. Pixel Watch 3 produced a strong current result against maximal laboratory testing, but the validation was Google-funded and has not yet been independently replicated across multiple models or cohorts.
- Laboratory September 15, 2026The Accuracy of VO2 max, Heart Rate and Energy Expenditure Measurements From Google Pixel Watch 3
Pixel Watch 3 VO2 max estimates showed useful group-level agreement but non-trivial individual error; reported MAPE was about 8.7%.
PolarPolar Vantage M2, Polar V8002 studies
Polar ranks fourth. The brand has substantial historical VO2 max research, but Polar V800 was clearly behind Garmin in a direct comparison and the newer Vantage M2 resting estimate also showed larger error.
- Clinical September 10, 2025Validation of the Polar Fitness Test for Estimation of Maximal Oxygen Consumption at Rest in Medically Supervised Exercise Training: Comparison With CPET and the 6-Minute Walk Test
Polar Fitness Test VO2max showed moderate agreement, with MAPE around 13.7% and ICC around 0.74.
- Laboratory August 22, 2019Validity of Wrist-Worn Activity Trackers for Estimating VO2max and Energy Expenditure
Polar V800 had MAPE 13.2%, MAE 6.5 mL/kg/min and ICC 0.67. Mean VO2max was about 3.0 mL/kg/min above spirometry, and error was substantially larger than Garmin Forerunner 920XT in the same participants.
AppleApple Watch Series 10, Apple Watch, Apple Watch Series 73 studies
Apple ranks fifth. Series 7, Series 10 and broader Apple Watch studies provide unusually current independent evidence, but error has stayed around the low-to-mid teens and individual limits of agreement remain wide.
- Laboratory April 1, 2026Accuracy of VO2 max Estimates From Apple Watch Series 10
Series 10 underestimated VO2max by 6.25 mL/kg/min on average; MAE was 6.79 and MAPE 13.2%.
- Laboratory May 15, 2025Investigating the Accuracy of Apple Watch VO2 max Measurements: A Validation Study
Apple Watch VO2max showed mean absolute percentage error around 13.3% in this validation study, with wide individual limits of agreement.
- Validation July 31, 2024Assessing the Accuracy of Smartwatch-Based Estimation of Maximum Oxygen Uptake Using the Apple Watch Series 7: Validation Study
Apple Watch Series 7 underestimated laboratory VO2max by 4.51 mL/kg/min on average. MAPE was 15.79% overall and error increased markedly in participants with excellent fitness.
WHOOPWHOOP 5.0 / WHOOP MG1 study
WHOOP remains unranked. Its internal laboratory validation is encouraging, especially for the GPS-augmented model, but manufacturer testing is not weighted like independent peer-reviewed validation.
- Laboratory 2026How Accurate Is WHOOP VO2 Max? The Science Behind the Estimate
WHOOP reported MAPE 8.0% for its passive model and 7.1% for its GPS-augmented model, with mean absolute error 3.7 and 3.3 mL/kg/min respectively. The validation was manufacturer-run and has not been independently replicated.