What Is a Good GMI Range & How to Improve It?

A good GMI for most adults with diabetes falls below 7.0%, which corresponds to an average sensor glucose of roughly 154 mg/dL. That target mirrors the widely used HbA1c goal of less than 7.0%, and for good reason: GMI was specifically designed to estimate HbA1c from continuous glucose monitor (CGM) data. But the two numbers do not always match, and knowing why they drift apart matters just as much as knowing the target itself.

What GMI Actually Tells You

GMI stands for Glucose Management Indicator. It replaced an older metric called “estimated A1C” (eA1C) after researchers recognized that the original formula and the confusing name were causing problems. The updated formula was built from recent clinical trials using the most accurate CGM systems available at the time, and the name change was meant to signal that this is a CGM-derived estimate, not a laboratory measurement of hemoglobin glycation.¹ The calculation takes your average sensor glucose over a period (typically 14 days or more) and converts it into a percentage that looks like an HbA1c value. If your CGM shows an average glucose of about 154 mg/dL, your GMI will be around 7.0%. Lower average glucose means a lower GMI.

Because it updates with every new day of CGM data, GMI gives you a much more responsive picture of glucose control than a lab draw you get every three months. You can watch it shift in near-real time as you change meals, activity, or medication. That responsiveness is its biggest advantage and, as we will see, a source of some confusion.

Why GMI and HbA1c Often Disagree

Your lab HbA1c reflects how much glucose has attached to hemoglobin in your red blood cells over the last two to three months. GMI, on the other hand, only knows what the sensor measured. Those two numbers can diverge for reasons that have nothing to do with how well you are managing your glucose.

One major category is biology. People glycate hemoglobin at different rates. If your body happens to glycate faster than average, your lab HbA1c will read higher than your CGM-derived GMI, even though your actual glucose levels are the same as someone whose numbers match. Conditions like anemia, chronic kidney disease, and hemoglobin variants also change how long red blood cells circulate or how hemoglobin behaves in the lab assay, pushing HbA1c up or down independent of glucose.² A study in patients with chronic kidney disease and anemia found that HbA1c systematically overestimated glucose control compared to GMI, with an average gap of about 0.76 percentage points between the two.³

Another category is timing. HbA1c integrates roughly 90 days, weighting the most recent weeks more heavily. GMI can be calculated from as little as 14 days. If you had a rough month of glucose control two months ago but have since improved, your GMI will reflect the improvement sooner than your lab value will. The reverse also happens: a few recent weeks of poor control will move GMI quickly while HbA1c lags behind. Researchers studying discordance between the two measures often exclude patients with known anemia, kidney disease, liver disease, or other conditions precisely because those conditions make the comparison unreliable.⁴

If your GMI consistently runs lower than your lab HbA1c by more than about half a percentage point, it is worth discussing with your doctor. One of the two is giving you an overly optimistic or pessimistic picture, and the practical consequence is that treatment decisions based solely on the wrong one could miss the mark.

GMI, Time in Range, and Glucose Variability

GMI is just an average, and averages can hide important patterns. Two people can both have a GMI of 7.0%, but one stays between 130 and 180 mg/dL most of the day while the other swings between 60 and 300 mg/dL. The second person has much more glucose variability and spends far more time dangerously high or low, despite the identical average.

That is why clinicians increasingly look at time in range (TIR), typically defined as the percentage of the day spent between 70 and 180 mg/dL. GMI and TIR are strongly correlated. In pregnant women with type 1 diabetes, for example, the correlation between TIR and GMI was about 0.9 in each trimester.⁵ But glucose variability complicates the relationship. In adults with type 1 diabetes, a study found that when glucose variability (measured by coefficient of variation) was within a healthy range, a GMI of 7.0% predicted a TIR about eight percentage points higher than when variability was excessive.⁶ In other words, if your GMI looks decent but your glucose swings are wide, your actual time in a safe range is worse than the GMI alone would suggest.

The practical takeaway: aim for a GMI below 7.0%, but also pay attention to your TIR (a common goal is above 70% for most adults with diabetes) and your coefficient of variation (ideally below 36%). A low GMI with high variability is not the same as a low GMI with smooth, stable glucose.

How Diet Influences GMI

Since GMI is calculated from average sensor glucose, anything that lowers your daily glucose readings will lower your GMI. Diet is the most direct lever.

Carbohydrate-restricted diets have the most studied effect on CGM metrics. A meta-analysis of trials in people with type 2 diabetes found that carb-restricted diets significantly reduced 24-hour mean blood glucose, with a moderate effect size, and exploratory analysis suggested the benefit grew with longer adherence.⁷ Lower mean glucose directly translates to lower GMI.

That does not mean extremely low carb intake is always better. A CGM-based study in people with type 2 diabetes found that those consuming a moderate carbohydrate intake of 55 to 60% of total calories had an average GMI of 6.6%, while those eating more than 60% of calories from carbohydrates had an average GMI of 7.06%.⁸ The sweet spot, at least in this population, was moderate rather than extreme restriction. Whole food quality, fiber content, and meal composition all influence how sharply glucose rises after eating, and those spikes contribute to the average that GMI reflects.

Meal Order Can Blunt Glucose Spikes

One surprisingly effective and easy-to-implement strategy involves the order in which you eat the components of a meal. Eating protein or fat before carbohydrates promotes the release of gut hormones that slow stomach emptying and improve insulin response, which flattens the postprandial glucose curve.⁹

Recent CGM-based research confirmed this, showing that consuming fiber and protein before carbohydrates significantly reduced postprandial glucose rises compared to eating carbohydrates alone. The effect was most evident in the first two hours after eating and was strongest when fiber and protein were combined before the carbohydrate portion. The benefit appeared in both healthy people and those with type 2 diabetes.¹⁰ In practical terms, eating your salad and chicken before the rice or bread at dinner can meaningfully lower the glucose spike from that meal, and those smaller spikes add up to a lower 24-hour average and, eventually, a lower GMI.

Exercise Lowers Average Glucose Beyond the Workout Itself

Physical activity improves insulin sensitivity and directly draws glucose out of the bloodstream for fuel, both of which lower average sensor glucose. What is less well known is that the benefit extends well past the workout itself. In a crossover study of people with type 1 diabetes, mean interstitial glucose was lower on exercise days compared to rest days, and when exercise was performed in the evening, glucose levels were also significantly lower the following day.¹¹ That next-day carry-over effect means even a few weekly sessions of moderate activity can pull your 14-day average glucose down enough to move the GMI needle.

You do not need marathon-level training. Walking after meals, resistance training a few times a week, or any movement that gets your heart rate up consistently tends to bring average glucose down. The key is consistency rather than intensity, because GMI reflects every day in its calculation window, and rest days with higher glucose pull the average back up.

How Sleep and Stress Push GMI Up

Poor sleep is one of the most underappreciated contributors to elevated glucose. In healthy young adults, just three days of restricted sleep produced a roughly 5% increase in 24-hour glucose levels measured by CGM, with both overnight and daytime glucose rising significantly.¹² A 5% bump in mean glucose may not sound dramatic, but sustained over weeks, it adds up.

For people with type 2 diabetes, the consequences of bad sleep appear to go further. Poor sleep quality has been linked to a more pronounced dawn phenomenon, the early-morning spike in glucose that many people with diabetes experience. Researchers found that people with type 2 diabetes who slept poorly had higher dawn glucose rises and disrupted expression of circadian clock genes.¹³ Since the dawn phenomenon can push fasting glucose up for several hours each morning, it has an outsized effect on your daily average and therefore your GMI.

Stress operates through similar pathways. Cortisol and other stress hormones raise blood glucose directly and impair insulin sensitivity. If you are doing everything right with diet and exercise but your GMI is stubbornly high, sleep quality and chronic stress are worth evaluating.

Medications That Move GMI

When lifestyle changes are not enough, medications can substantially lower average glucose and, by extension, GMI. The specific drug class matters less than matching the right therapy to your situation, but some deserve mention for their particularly strong effect on CGM-visible metrics.

GLP-1 receptor agonists like semaglutide, liraglutide, and dulaglutide are among the most effective at reducing HbA1c and, by implication, GMI. Long-acting versions of these drugs have especially strong effects on overnight and fasting glucose, which are major contributors to the 24-hour average. They are now recommended as the preferred first injectable therapy for type 2 diabetes, ahead of insulin, because of their similar or superior glucose-lowering effect combined with weight loss and a low risk of causing hypoglycemia.⁸ SGLT2 inhibitors, metformin, and insulin itself are all standard tools as well. The choice depends on your diabetes type, kidney function, cardiovascular risk profile, and other individual factors.

If you are on a CGM, medications give you an unusually clear feedback loop. You can see the effect of a new drug or dose adjustment reflected in your GMI within about two weeks, rather than waiting three months for a follow-up lab.

How Many Days of CGM Data Do You Need?

A GMI calculated from three days of CGM data is not reliable. The number shifts meaningfully depending on how much data goes into it. Research consistently points to 14 days as the minimum useful window. In a study of more than 800 adults with type 2 diabetes not using insulin, GMI from 14 days of data correlated strongly with 90-day values, and extending to 30 days only slightly improved accuracy.¹⁴

Sensor wear percentage also matters. If you wear your CGM only half the time, you are missing data from hours that may look very different from the hours you captured. An analysis of suboptimal CGM use found that even when sensor wear ranged from 45% to 95% over 90 days, 14 days was sufficient for mean glucose and time-in-range estimates, though metrics like glucose variability needed at least 28 days to stabilize.¹⁵ One study used a combined approach of up to 28 days of CGM data from two different sensor brands to calculate more robust metrics.¹⁶

For day-to-day self-management, the rule of thumb is simple: wear your sensor consistently for at least two weeks before drawing any conclusions from your GMI. If you have gaps in wear, you need more data, not less, to get a trustworthy number.

When Medical Conditions Skew the Numbers

Certain health conditions make the GMI-to-HbA1c comparison particularly unreliable, which can cause confusion if you are tracking both.

Kidney disease and anemia are the most common culprits. Anemia shortens the lifespan of red blood cells, which can artificially lower or raise HbA1c depending on the type of anemia. Chronic kidney disease introduces its own hemoglobin abnormalities. In patients with both conditions, one pilot study found that HbA1c overestimated glucose control compared to what the CGM data showed, with poor agreement between the two values across all stages of kidney disease severity studied.³ For these patients, GMI from a CGM may actually give a more accurate picture of daily glucose control than the lab HbA1c, since it bypasses the hemoglobin entirely.

Hemoglobin variants, pregnancy (which changes blood volume and red cell turnover), recent blood transfusions, and certain medications like erythropoietin-stimulating agents can also throw off HbA1c without affecting GMI. If you have any of these conditions, discuss with your care team which metric deserves more weight in treatment decisions.

Sensor Accuracy and What GMI Cannot See

GMI is only as good as the sensor data it is built from. Modern CGM sensors are quite accurate, but they measure glucose in the interstitial fluid beneath the skin rather than directly in the blood. There is a small time lag between what is happening in your blood and what the sensor reads. For the Dexcom G7, that lag averages about 3.5 minutes.¹⁷ That is short enough to be clinically irrelevant for most purposes, including GMI calculation over days or weeks.

Where sensor limitations matter more is in the first day of a new sensor session, when readings can be less stable, and during periods of rapid glucose change, such as immediately after a high-carbohydrate meal or intense exercise. These inaccuracies tend to wash out over a 14-day window, but they can make hour-by-hour data look noisier than the underlying truth. Compression lows, caused by sleeping on the sensor, are another common artifact that can pull your average down falsely. If your overnight glucose regularly dips to improbable lows around 3-4 AM, suspect compression before assuming real hypoglycemia.

GMI in People Without Diabetes

CGMs are increasingly popular among people without diabetes who want to “optimize” their metabolic health. If you fall into this group, be cautious about interpreting your GMI through the lens of diabetes targets.

The GMI formula was developed and validated in populations with diabetes. When applied to people with normal glucose levels, it tends to overestimate what their actual lab HbA1c would be. A recent analysis found that among people with a true HbA1c below 5.7% (the threshold for prediabetes), only 19% had a standard GMI that also fell below 5.7%. An updated version of the formula improved this to 63%, but that still means more than a third of genuinely normoglycemic people would be incorrectly flagged as prediabetic by their CGM’s GMI reading.¹⁸

This overestimation can cause unnecessary anxiety. If you are wearing a CGM for general wellness and your GMI reads 5.8% or 5.9%, do not assume you have prediabetes. A fasting glucose test or an actual HbA1c lab draw is a much more reliable screen. The current GMI formula was simply not designed for the normoglycemic range, and using it there introduces systematic error that inflates the result.

Putting It All Together Without Obsessing

CGM data can be empowering or overwhelming depending on how you use it. Checking your GMI every few hours is not productive because the number barely moves day to day. A more practical rhythm is to review your ambulatory glucose profile and GMI every two weeks, paying attention to trends rather than single-day swings. If your GMI has drifted up by a few tenths of a percentage point over consecutive two-week windows, that is a signal worth acting on. If it bounced up for one chaotic week and came right back down, it probably does not warrant a medication change.

The emerging landscape of digital diabetes tools, including connected insulin pens and integrated CGM-pump systems, promises to make glucose management more automated. But research on the efficacy and safety of many of these newer devices is still limited, and their cost-effectiveness is only starting to be studied in clinical settings. For now, the fundamentals remain what they have always been: consistent CGM wear for reliable data, moderate dietary changes sustained over time, regular movement, decent sleep, and honest conversations with your care team about what the numbers mean and which ones to trust.