The Glucose Management Indicator (GMI) is calculated with a simple formula: GMI (%) = 3.31 + 0.02392 × your mean glucose in mg/dL. All you need is your average glucose reading from a continuous glucose monitor (CGM) over at least 14 days. If your mean glucose is 180 mg/dL, for example, your GMI would be 7.6%.
GMI gives you an estimate that looks like an A1C percentage, but it’s derived entirely from CGM data rather than a blood draw. Understanding how it works, what the numbers mean, and why it sometimes disagrees with your lab A1C can help you get more out of your glucose data.
The GMI Formula
The equation was published in Diabetes Care in 2018 and has since become the standard used by CGM software and clinic reports. In the units most commonly used in the United States:
GMI (%) = 3.31 + 0.02392 × [mean glucose in mg/dL]
If your CGM software reports glucose in mmol/L (common outside the U.S.), the equivalent formula is:
GMI (mmol/mol) = 12.71 + 4.70587 × [mean glucose in mmol/L]
The “mean glucose” here is the simple average of all your CGM readings over the measurement period. Most CGM apps calculate this for you automatically on your summary report, but you can also find it by exporting your data and averaging the values yourself.
A Step-by-Step Example
Say your CGM report shows a mean glucose of 160 mg/dL over the past two weeks. Plug that into the formula:
- Multiply 160 by 0.02392 = 3.83
- Add 3.31 = 7.14
- Your GMI is approximately 7.1%
That’s it. The math is straightforward once you have a reliable mean glucose number. For quick reference, here are several common mean glucose values and their corresponding GMI results:
- 126 mg/dL → GMI 6.3%
- 140 mg/dL → GMI 6.7%
- 150 mg/dL → GMI 6.9%
- 160 mg/dL → GMI 7.1%
- 170 mg/dL → GMI 7.4%
- 180 mg/dL → GMI 7.6%
- 190 mg/dL → GMI 7.9%
- 200 mg/dL → GMI 8.1%
Every 10 mg/dL increase in mean glucose raises GMI by roughly 0.24 percentage points.
Data Requirements for a Valid GMI
A single day of CGM readings won’t produce a meaningful GMI. The calculation requires a minimum of 14 days of CGM data to capture enough glucose variability across meals, sleep, activity, and stress. On top of that, your sensor should be actively recording for at least 70% of each day. Gaps from sensor warmup periods, lost signal, or days without wearing the device can undermine the accuracy of the average.
Most CGM platforms (Dexcom Clarity, LibreView, and others) will display GMI on your summary report once you meet these thresholds. If your wear time falls below the minimum, the report may omit GMI or flag it as unreliable.
How GMI Differs From Lab A1C
GMI is designed to approximate A1C, but they measure fundamentally different things. A1C reflects how much sugar has attached to your red blood cells over roughly two to three months. GMI, on the other hand, is a mathematical conversion of your average sensor glucose. The two numbers frequently don’t match, and neither one is “wrong” when they diverge.
The most common reason for a mismatch is red blood cell lifespan. A1C assumes your red blood cells live about 120 days on average, but this varies from person to person. If your red blood cells turn over faster than average, your lab A1C will read lower than your GMI suggests. If they last longer, A1C will read higher. Research has shown that normal variation in red cell lifespan among otherwise healthy people is enough to shift A1C results.
Certain medications also create discordance. SGLT2 inhibitors, a class of diabetes drugs that work by causing the kidneys to excrete excess glucose, have been shown to alter the relationship between A1C and CGM-derived glucose in a way that makes A1C appear disproportionately low relative to actual glucose levels. Other factors that can widen the gap include anemia, iron deficiency, kidney disease, and pregnancy.
The American Diabetes Association’s 2025 Standards of Care describe GMI as a “calculated value approximating A1C” while noting it is “not always equivalent.” When the two numbers disagree by more than about 0.5 percentage points, the gap itself becomes useful information: it signals that something beyond average glucose is influencing one of the metrics.
Why GMI Is Useful Despite Its Limits
If GMI isn’t always the same as A1C, you might wonder why it exists. The main advantage is timing. A1C is a backward-looking average weighted toward the most recent four to six weeks of glucose exposure, and you typically get it checked every three to six months. GMI updates every time your CGM generates a new 14-day report. That means you can see whether a medication change, dietary shift, or exercise routine is moving the needle weeks before your next lab draw.
GMI also eliminates the biological noise that makes A1C unreliable for some people. If you have a hemoglobin variant, chronic kidney disease, or any condition that affects red blood cell turnover, your lab A1C may never accurately reflect your glucose control. GMI bypasses red blood cells entirely and reports what your sensor actually measured.
The flip side is that GMI only captures what the sensor sees. It can’t account for glucose patterns during periods you weren’t wearing the device, and it treats all glucose readings equally. A person with stable glucose at 160 mg/dL all day and a person who swings between 80 and 240 mg/dL could have the same mean glucose and the same GMI, despite very different glucose profiles. That’s why clinicians look at GMI alongside time in range, time below range, and glucose variability metrics rather than relying on any single number.
Doing the Calculation Yourself
In practice, most people never need to do the math by hand. Dexcom Clarity, Abbott’s LibreView, and other CGM reporting platforms calculate GMI automatically once you have enough sensor data. The number appears on your Ambulatory Glucose Profile (AGP) report, which is the standardized one-page summary your healthcare provider reviews.
If you do want to calculate it yourself, export your CGM data to a spreadsheet, average all glucose readings from a 14-day window (or longer), and apply the formula. A longer window, such as 30 or 90 days, will smooth out short-term fluctuations and give you a more stable estimate. Just make sure you had consistent sensor wear throughout the period. A 30-day average where you only wore the sensor for 10 of those days will be skewed toward whichever days happened to have data.