Diabetes Tech Clinic
Home CGMs

Continuous Glucose Monitors (CGMs)

Compare specs, accuracy (MARD), and app features of the leading CGM systems available in 2024.

Continuous Glucose Monitors (CGMs) have replaced routine fingersticks as the standard of care for intensive insulin management. The current generation of sensors are smaller, require no fingerstick calibrations, and stream data directly to smartphones and insulin pumps.

Latest Generation CGMs

Dexcom G7

10-Day Wear

Dexcom's smallest sensor yet. All-in-one disposable design with a 30-minute warmup and a 12-hour grace period at the end of the session.

MARD (Adults) 8.2%
Warmup Time 30 mins
Pump Integration Tandem, iLet (Omnipod rolling out)
Read Full Review

FreeStyle Libre 3

14-Day Wear

Abbott's tiny sensor that streams real-time data continuously without needing to scan. The smallest and thinnest CGM currently available.

MARD (Adults) 7.9%
Warmup Time 60 mins
Pump Integration Tandem Mobi (Select markets)
Read Full Review

Can't decide?

Read our detailed head-to-head comparison analyzing accuracy, app experience, alarms, and overall cost.

Compare Dexcom G7 vs Libre 3

Previous Generation (Still Widely Used)

Dexcom G6

10-Day Wear

The gold standard for AID integration. While replaced by the G7, the G6 remains the most widely integrated sensor for closed-loop pump systems including Omnipod 5 and Tandem t:slim X2. Features a separate sensor wire and reusable transmitter (90 days).

MARD: 9.0%
2-hr Warmup

Understanding MARD (Mean Absolute Relative Difference)

MARD is the standard metric used to express CGM accuracy. It represents the average error between the CGM's reading and a simultaneous reference blood glucose measurement (usually YSI in clinical trials).

A lower MARD percentage indicates better accuracy. Generally, a MARD below 10% is considered highly accurate and sufficient for making insulin dosing decisions without a fingerstick confirmation. However, MARD can vary based on:

  • Day of wear (often less accurate on day 1)
  • Glucose trends (accuracy often drops during rapid changes)
  • Glucose ranges (often less accurate in extreme hypoglycemia)