GMI Calculator (Glucose Management Indicator)
Medically reviewed by Dr. Chhaya Makhija, MD, DipABLM — September 2026
Turn your CGM average glucose into your glucose management indicator (GMI) — the modern, CGM-based estimate of your A1c-equivalent, in both % and mmol/mol.
Your mean sensor glucose, ideally over 14+ days.
The single most useful thing to understand about GMI is what it is not: it is not a prediction of your next A1c.
It looks like one. It is expressed as a percentage, it lands in the same range, and CGM apps display it where an A1c would go. But GMI is calculated from your sensor's mean glucose over roughly the last two weeks, while A1c measures glycated haemoglobin accumulated over eight to twelve. They are different measurements of different things over different periods.
That is why the two so often disagree, and why the disagreement is not evidence that either is broken. A gap of up to about half a percentage point in either direction is ordinary. Red blood cell lifespan varies between people, and A1c quietly assumes an average that many individuals do not have — while GMI is blind to haemoglobin entirely and instead inherits whatever bias your sensor carries.
Where GMI genuinely wins is speed. It responds within two weeks of a change in management, where A1c takes months to catch up.
Enter your CGM's average glucose below. Beneath the calculator you will find the formula, a full lookup table, how GMI appears on an AGP report, and how to read a persistent gap between GMI and your lab result.
How to Use This Calculator
- Choose your glucose units — mg/dL or mmol/L.
- Enter your CGM average (mean) glucose, ideally from at least 14 days of data with good sensor wear.
- Read your GMI in both % and mmol/mol.
GMI estimates your A1c from CGM data; it can differ from a lab A1c by 0.3% or more, so use it alongside — not instead of — your clinician's testing.
How GMI Is Calculated
The glucose management indicator replaced the older "estimated A1c" label. It is calculated directly from your mean CGM glucose using a validated regression equation, giving a number on the same scale as A1c:
Example: a mean glucose of 154 mg/dL (8.5 mmol/L) → 3.31 + 0.02392 × 154 = GMI ≈ 7.0%.
GMI by Average Glucose
| Mean glucose (mg/dL) | GMI (%) |
|---|---|
| 126 | 6.3 |
| 154 | 7.0 |
| 183 | 7.7 |
| 212 | 8.4 |
| 240 | 9.1 |
Use at least 14 days of CGM data (ideally with >70% sensor wear) for a reliable GMI.
GMI on Your CGM Report (AGP) & Target Goals
The standardized AGP report
GMI appears at the top of the Ambulatory Glucose Profile (AGP) — the one-page summary your CGM software produces (Dexcom Clarity, LibreView, Tidepool). Alongside GMI it shows your average glucose, time in range, and glucose variability, giving a fuller picture than a single number.
Recommended CGM targets
International consensus targets for most adults with type 1 or type 2 diabetes are:
| Metric | Goal |
|---|---|
| Time in range (70–180 mg/dL; 3.9–10.0 mmol/L) | > 70% |
| Time below range (< 70 mg/dL; 3.9 mmol/L) | < 4% (< 1% below 54) |
| Time above range (> 180 mg/dL; 10.0 mmol/L) | < 25% |
| GMI | < 7% |
| Glucose variability (CV) | ≤ 36% |
Targets are looser for older or higher-risk people and tighter in pregnancy. Your care team sets your individual goals.
GMI Lookup Table: Mean Glucose → GMI
Computed from the published formula GMI (%) = 3.31 + 0.02392 × mean glucose (mg/dL) (Bergenstal et al., Diabetes Care 2018).
| Mean glucose (mg/dL) | Mean glucose (mmol/L) | GMI (%) | GMI (mmol/mol) |
|---|---|---|---|
| 100 | 5.6 | 5.7 | 39 |
| 114 | 6.3 | 6.0 | 42 |
| 125 | 6.9 | 6.3 | 45 |
| 140 | 7.8 | 6.7 | 50 |
| 154 | 8.6 | 7.0 | 53 |
| 167 | 9.3 | 7.3 | 56 |
| 180 | 10.0 | 7.6 | 60 |
| 200 | 11.1 | 8.1 | 65 |
| 220 | 12.2 | 8.6 | 70 |
| 240 | 13.3 | 9.1 | 76 |
| 260 | 14.4 | 9.5 | 80 |
| 280 | 15.5 | 10.0 | 86 |
| 300 | 16.7 | 10.5 | 91 |
The often-quoted anchor: a 14-day mean of 154 mg/dL (8.5 mmol/L) ≈ GMI 7.0%. For the reverse direction (A1c → average glucose), use the HbA1c / eAG calculator.
Your GMI Number: What It Means
Most people arrive here holding a GMI already. The figure sits on a Dexcom Clarity, LibreView or AGP report, and the question is what it represents. This table runs the formula backwards — start in the first column and read across.
| Your GMI (%) | Mean glucose (mg/dL) | Mean glucose (mmol/L) | GMI (mmol/mol) | Typical lab A1c span |
|---|---|---|---|---|
| 5.7 | 100 | 5.6 | 39 | 5.2 – 6.2 |
| 6.0 | 112 | 6.2 | 42 | 5.5 – 6.5 |
| 6.5 | 133 | 7.4 | 48 | 6.0 – 7.0 |
| 6.6 | 138 | 7.7 | 49 | 6.1 – 7.1 |
| 7.0 | 154 | 8.6 | 53 | 6.5 – 7.5 |
| 7.2 | 163 | 9.1 | 55 | 6.7 – 7.7 |
| 7.5 | 175 | 9.7 | 58 | 7.0 – 8.0 |
| 7.8 | 188 | 10.4 | 62 | 7.3 – 8.3 |
| 8.0 | 196 | 10.9 | 64 | 7.5 – 8.5 |
| 8.5 | 217 | 12.1 | 69 | 8.0 – 9.0 |
| 9.0 | 238 | 13.2 | 75 | 8.5 – 9.5 |
Mean glucose is derived from GMI = 3.31 + 0.02392 × mg/dL, solved for glucose. The final column is a span, not a conversion — see the caution below.
Why the last column is a range and not a number
Searches for "GMI to A1c" assume a fixed conversion exists. It does not, and the term GMI was introduced in 2018 precisely to retire the older label "estimated A1c". Your GMI describes sensor glucose over 14 days. Your lab A1c describes how much glucose attached to your haemoglobin over roughly three months. Two different measurements over two different windows produce two different numbers.
In the original Bergenstal dataset only about half of participants landed within 0.3% of their lab A1c, and a minority differed by more than a full percentage point. The span above covers roughly GMI ±0.5%, which captures most people but not everyone. Treat it as the neighbourhood your lab result probably falls in, then use your own measured gap once you have both numbers side by side.
Reading three common values
Is a GMI of 6.6 good? For most non-pregnant adults with diabetes, yes. It reflects a 14-day average near 138 mg/dL (7.7 mmol/L) and sits below the widely used A1c target of 7%. Pair it with time in range: the standard goal is above 70% between 70 and 180 mg/dL, and GMI alone hides how much you swing around that average.
What does a GMI of 7.2 mean? Your sensor averaged about 163 mg/dL (9.1 mmol/L). That sits modestly above the common 7% target, so most care teams treat it as a nudge rather than an alarm — often a basal or ratio adjustment rather than a regimen change.
What about 7.8? The 14-day average runs near 188 mg/dL (10.4 mmol/L), and a lab A1c drawn the same week most often returns somewhere between 7.3 and 8.3. This is the band where teams usually look at overnight patterns and post-meal spikes before changing total dose.
Targets are individual. Older adults, people with hypoglycaemia unawareness, and people in pregnancy all work to different numbers. Confirm your own target with your diabetes care team rather than reading it off a table.
When GMI and Lab A1c Disagree
In the original Bergenstal dataset, only about half of people had a GMI within 0.3% of their lab A1c — disagreement is normal, not an error. Persistent gaps come from real physiology:
- Red-cell lifespan: longer-lived red cells accumulate more glycation → lab A1c reads higher than GMI; shorter lifespan (anemia, hemolysis, kidney disease, recent blood loss) → A1c reads lower.
- Glycation rate variability: individuals glycate hemoglobin at different rates at the same glucose level.
- Hemoglobin variants & pregnancy can distort the A1c assay itself.
- Sensor bias or poor wear time (<70% over 14 days) can distort the GMI side instead.
Clinically, a stable personal gap is useful information: if your GMI always runs ~0.4% below your lab A1c, your team can calibrate targets around your pattern. A gap that suddenly changes is worth investigating (new anemia, kidney change, sensor issue).
Limitations & Safety Notes
- GMI is calculated from CGM mean glucose. It is not a haemoglobin measurement and is not expected to equal your lab A1c — a gap between the two is normal, not an error.
- It is only meaningful with adequate CGM data, conventionally at least 14 days with 70% or more sensor wear.
- Anything that changes red-blood-cell lifespan — anaemia, recent transfusion, pregnancy, chronic kidney disease, haemoglobin variants — widens the gap between GMI and measured A1c.
- A consistent, repeatable difference between your GMI and your A1c is clinically useful information and worth showing your care team rather than explaining away.
Frequently Asked Questions
What is GMI and how is it different from A1c?
GMI (glucose management indicator) estimates your A1c from CGM average glucose over a set period. A1c is a blood test reflecting roughly 2–3 months of glucose. GMI and A1c often differ slightly because they measure in different ways.
Why is my GMI different from my lab A1c?
Individual biology — red blood cell lifespan and how glucose attaches to hemoglobin — varies. So two people with the same GMI can have different lab A1c values. A persistent gap is worth discussing with your clinician.
How much CGM data do I need for an accurate GMI?
At least 14 days of CGM data with good sensor wear (over about 70% of the time) gives a representative GMI. Shorter or patchy data can skew the average glucose and the resulting GMI.
Is GMI the same as estimated average glucose (eAG)?
No. GMI converts average glucose into an A1c-like percentage. eAG goes the other way — it converts an A1c into an average glucose. See our eAG ↔ A1c calculator for that conversion.
What is a good GMI?
For most adults with diabetes, a GMI under 7% is the common target, mirroring the usual A1c goal. Goals are individualized — looser for older or higher-risk people and tighter in pregnancy — so confirm your personal target with your care team and read GMI alongside your time in range.
Where do I find my GMI?
Your CGM software calculates it automatically and shows it on the Ambulatory Glucose Profile (AGP) report — in apps such as Dexcom Clarity, LibreView or Tidepool. If you only have your average glucose, this calculator gives you the same number.
Is a GMI of 6.6 good?
For most non-pregnant adults with diabetes, yes. A GMI of 6.6% reflects a 14-day average glucose near 138 mg/dL (7.7 mmol/L), which sits below the widely used A1c target of 7%. Read it alongside time in range rather than on its own: the standard goal is more than 70% of readings between 70 and 180 mg/dL, and two people with the same GMI can have very different swings around that average. Targets differ for older adults, for people with hypoglycaemia unawareness and in pregnancy, so confirm your own target with your diabetes care team.
What is the A1c equivalent of a 7.8 GMI?
There is no fixed equivalent, and that is the point of the term GMI. A GMI of 7.8% corresponds to a 14-day sensor average near 188 mg/dL (10.4 mmol/L). A laboratory A1c drawn the same week most often returns between roughly 7.3% and 8.3%, because GMI measures sensor glucose over 14 days while A1c measures haemoglobin glycation over about three months. In the original Bergenstal dataset only about half of people fell within 0.3% of their lab A1c, and a minority differed by more than a full percentage point. Once you have both numbers, your own personal gap is more useful than any published conversion.
Sources
- Bergenstal RM, et al. Glucose Management Indicator (GMI): A new term for estimating A1C from CGM. Diabetes Care. 2018. Link
- American Diabetes Association. Standards of Care — CGM metrics and time in range. Link
Written by Ryan Mitchell · Last updated: September 2026