GMI Calculator (Glucose Management Indicator)
Turn the average glucose from your CGM into a Glucose Management Indicator (GMI) — the same figure your sensor report uses to estimate where a laboratory A1C might land. Enter the mean glucose from your CGM app in mg/dL or mmol/L and the GMI updates as you type, with the formula and its source shown below.
Already have a GMI from your CGM report? Look it up in the GMI chart to see the CGM average glucose behind it, in mg/dL and mmol/L.
Calculate
Enter your CGM average glucose
Use the mean glucose from your CGM app and pick its unit. For a meaningful GMI, the average should cover at least 14 days of sensor data — a shorter stretch describes those days, not your usual glucose.
Glucose Management Indicator (GMI)
Enter your CGM average above to see the GMI.
Look up
GMI chart: the CGM average behind each GMI value
Find your GMI in the left column: the two columns to its right show the CGM average glucose that produces it, in mg/dL and mmol/L. GMI is reported on the same percentage scale as A1C. Each row runs the same Bergenstal 2018 formula the calculator above uses, solved exactly in reverse — mean glucose (mg/dL) = (GMI − 3.31) ÷ 0.02392 — with mg/dL rounded to a whole number and mmol/L to one decimal. The chart only converts numbers; it does not grade any value.
| GMI | CGM mean glucose (mg/dL) | CGM mean glucose (mmol/L) |
|---|---|---|
| 5.5% | 92 | 5.1 |
| 5.6% | 96 | 5.3 |
| 5.7% | 100 | 5.5 |
| 5.8% | 104 | 5.8 |
| 5.9% | 108 | 6.0 |
| 6.0% | 112 | 6.2 |
| 6.1% | 117 | 6.5 |
| 6.2% | 121 | 6.7 |
| 6.3% | 125 | 6.9 |
| 6.4% | 129 | 7.2 |
| 6.5% | 133 | 7.4 |
| 6.6% | 138 | 7.6 |
| 6.7% | 142 | 7.9 |
| 6.8% | 146 | 8.1 |
| 6.9% | 150 | 8.3 |
| 7.0% | 154 | 8.6 |
| 7.1% | 158 | 8.8 |
| 7.2% | 163 | 9.0 |
| 7.3% | 167 | 9.3 |
| 7.4% | 171 | 9.5 |
| 7.5% | 175 | 9.7 |
| 7.6% | 179 | 10.0 |
| 7.7% | 184 | 10.2 |
| 7.8% | 188 | 10.4 |
| 7.9% | 192 | 10.6 |
| 8.0% | 196 | 10.9 |
| 8.5% | 217 | 12.0 |
| 9.0% | 238 | 13.2 |
| 9.5% | 259 | 14.4 |
| 10.0% | 280 | 15.5 |
This chart is for GMI values only. A same-period laboratory A1C commonly differs from GMI by around half a percentage point in either direction, so a lab A1C cannot simply be looked up here. To turn a lab A1C into an estimated average glucose, use the A1C to eAG conversion instead.
The math
The GMI formula and where it comes from
GMI (%) = 3.31 + 0.02392 × mean glucose (mg/dL)
This is the published Glucose Management Indicator regression (Bergenstal et al., Diabetes Care 2018), fitted on paired CGM and laboratory A1C data from just over 500 adults in clinical trials. The authors introduced the name GMI specifically to retire the older label “estimated A1C”: calling the sensor-derived number an A1C invited people to treat it as interchangeable with the lab test, which it is not. The formula expects mean glucose in mg/dL, so a mmol/L input on this page is first multiplied by 18.018 and only then enters the equation; the result is displayed to one decimal place, rounded half-up.
GMI is deliberately simple — an intercept and a slope. It captures how average glucose and A1C relate across a study population, and nothing about you personally: your red blood cells, your sensor, or which days the sensor was worn. That is why the number below your CGM average is an orientation, not a substitute for blood work.
GMI vs A1C
Why GMI and your lab A1C usually disagree — and why that is fine
Expect a gap. GMI and a same-period laboratory A1C commonly differ by around ±0.5 percentage points, and for a sizeable share of people the gap is larger still. The two numbers answer different questions: GMI summarizes what your sensor saw over a few weeks, while A1C reflects glucose exposure over the full lifespan of your red blood cells — and that lifespan genuinely differs from person to person. Anemia, kidney disease, pregnancy, recent blood loss or transfusion, and hemoglobin variants each move A1C without moving your glucose; sensor bias and wear gaps move GMI without moving your A1C.
A mismatch between the two is therefore not a data error. It does not mean your sensor is failing, your lab made a mistake, or your numbers are untrustworthy — it usually just means your personal glucose-to-A1C relationship sits off the population average the formula was fitted on. Some clinicians even use a stable personal gap as useful information in itself.
What GMI cannot do: it cannot replace a laboratory A1C, and it must not be used to diagnose anything or to change treatment on its own. Decisions belong with your care team, who will read GMI alongside time in range, variability, and the rest of your CGM report — and alongside actual blood work.
FAQ
GMI questions
Why is my GMI different from my lab A1C?
Because they measure different things over different windows. GMI is arithmetic on your sensor's mean glucose from the last few weeks; a laboratory A1C measures how much glucose has attached to hemoglobin over the whole lifespan of your red blood cells, roughly three months. Red cell lifespan itself varies from person to person, and anemia, kidney disease, pregnancy, recent blood loss, and hemoglobin variants all shift A1C without changing your actual glucose. On top of that, CGM sensors carry some bias of their own and only see the days you wore them. Gaps of around half a percentage point in either direction are common and expected — a mismatch does not mean your sensor data or your lab result is wrong.
How many days of CGM data do I need for a reliable GMI?
Use at least 14 days. The international consensus on CGM metrics recommends a 14-day window with the sensor active about 70% of the time or more, because two weeks of reasonably complete data correlate well with a full three months of glucose exposure. A GMI computed from just a few days — or from a stretch with long sensor gaps, illness, travel, or a medication change — describes that unusual stretch, not your usual glucose. Most CGM apps show the mean glucose for a selectable date range; pick 14, 30, or 90 days and enter that mean here.
Can GMI replace a lab A1C test?
No. GMI is an estimate of where a laboratory A1C might land if your recent sensor average continued — it is not a measurement of anything in your blood. Diagnosis and formal monitoring rely on an NGSP-certified laboratory A1C, and treatment changes should be decided with your care team using the full CGM picture (time in range, variability, lows), not a single derived number. GMI is best used the way sensor reports use it: as context for understanding your data between lab visits.
Which number from my CGM app should I enter?
Enter the average (mean) glucose, not the median and not the GMI the app already shows. Apps like Dexcom Clarity, LibreView, and CareLink display mean glucose on their summary or AGP report pages — that is the input this formula expects, in mg/dL or mmol/L. If your app already reports a GMI for the same date range, this calculator should reproduce it almost exactly; a small difference usually means the two are looking at different date ranges.
What does GMI mean?
GMI stands for Glucose Management Indicator. It is calculated from continuous glucose monitor (CGM) data rather than measured by a blood test, and it uses the same percentage scale as A1C to estimate the laboratory A1C that corresponds to your sensor's average glucose. Many CGM apps and reports show it right next to your average glucose.
Can I convert my GMI to an A1C?
There is nothing to convert: GMI is already on the A1C scale. A GMI of 6.5% is the formula's estimate of a laboratory A1C of about 6.5%. What the formula cannot tell you is your actual lab result — a same-period laboratory A1C commonly differs from GMI by around half a percentage point in either direction, and only an A1C blood test measures A1C itself.
What average glucose does my GMI correspond to?
GMI is defined by a straight-line formula, so it can be solved exactly for the CGM average that produces it: mean glucose (mg/dL) = (GMI − 3.31) ÷ 0.02392, and dividing the result by 18.018 gives mmol/L. For example, a GMI of 6.5% corresponds to a CGM average of about 133 mg/dL (7.4 mmol/L). The GMI chart on this page applies the same calculation row by row, so you can look up your GMI without doing the math. This only works backwards from GMI — it does not turn a laboratory A1C into an average glucose; for that, use the A1C to eAG calculator.
Is there a "good" GMI number?
Not one that applies to everyone. Glucose targets are set individually and depend on factors such as age, pregnancy, how often someone experiences low glucose, and other health conditions, so this site does not label any GMI value as good or bad — the chart on this page only converts numbers. Clinicians usually read GMI together with time in range and the frequency of lows rather than on its own, and your care team can tell you what your own numbers mean.