What the A1C calculator is doing
Your A1C and your average blood sugar are the same information in two formats, and one equation converts between them: estimated average glucose in mg/dL = 28.7 × A1C − 46.7. That formula comes from the ADAG study, which put continuous glucose monitors on hundreds of people, measured their actual average glucose for months, and fit a line to how it tracked their A1C. The American Diabetes Association has published the result ever since as “eAG.”
Run it in either direction. An A1C of 7.0% is an average glucose of about 154 mg/dL. A CGM average of 180 mg/dL corresponds to an A1C near 7.9%. Glucose is shown in both mg/dL and mmol/L, because which one you see depends on which country’s lab printed your report.
The bands, and what they actually mean
| A1C | Estimated average glucose | What it’s called |
|---|---|---|
| 5.0% | ~97 mg/dL (5.4 mmol/L) | Normal |
| 5.7% | ~117 mg/dL (6.5 mmol/L) | Start of prediabetes |
| 6.5% | ~140 mg/dL (7.8 mmol/L) | Diabetes threshold |
| 7.0% | ~154 mg/dL (8.6 mmol/L) | Common treatment target |
| 8.0% | ~183 mg/dL (10.2 mmol/L) | Above target for most adults |
| 9.0% | ~212 mg/dL (11.7 mmol/L) | Well above target |
Two things worth separating here. The 6.5% line is a diagnostic threshold, the number a clinician uses to say whether someone has diabetes. The ~7% figure is a treatment target, and it’s individual: younger people with a long horizon and no hypoglycemia risk often aim tighter, while older adults or people on insulin may be given a looser target on purpose, because driving A1C down carries its own risks.
Why A1C works at all
Glucose sticks to hemoglobin, the protein in your red blood cells, and it stays stuck for the life of that cell, about three months. Measure what fraction of your hemoglobin is glycated and you get a read on how much sugar has been circulating over that window, without anyone having to test you hourly for ninety days.
That’s the strength and the weakness. The strength is that A1C can’t be gamed by one good week of eating. The weakness is that it’s an average that hides its own shape: someone whose glucose swings from 60 to 250 can post the same A1C as someone who sits calmly at 155 all day, and those two people are not in the same clinical situation. It’s also weighted toward the recent past, roughly half of your A1C reflects the last month.
And it can be wrong outright. Anything that changes how long red blood cells live changes A1C without changing your glucose: anemia, sickle cell trait or disease, recent blood loss or transfusion, pregnancy, advanced kidney disease, and some medications. If your A1C and your CGM tell different stories, that gap is information, not a rounding error, bring it to your clinician.
The GLP-1 section
This is the part most A1C calculators skip. If you’re taking a GLP-1 (or deciding whether to), the useful question isn’t just what your A1C is, it’s what it’s likely to become.
The pivotal trial programs give the honest ranges. In their type 2 diabetes trials, tirzepatide (Mounjaro, Zepbound) lowered A1C by roughly two points or more; injectable semaglutide (Ozempic, Wegovy) by about 1.4 to 1.8 points; dulaglutide (Trulicity) by about 1.1 to 1.6; liraglutide (Victoza, Saxenda) by about 1 to 1.5; oral semaglutide (Rybelsus) by about 1 to 1.4. Pick your medication in the tool and it applies that range to your own starting number, so the result reads as a sentence rather than a statistic.
Three caveats that matter more than the numbers:
- Trial averages come from people starting high. Most of those participants began around 8%. Starting at 6.2%, you have far less room, and you will not see a two-point drop: the body doesn’t push A1C into hypoglycemic territory just because the average says so.
- Individual response varies enormously. Averages describe groups. Some people respond far better than the trial mean; some barely respond and switch drugs.
- These are diabetes-trial figures. If you’re taking a GLP-1 for weight loss without diabetes, your A1C is probably already in the normal or prediabetes range, and the change you’d see is much smaller.
The overlap between the two effects is real, though: weight loss itself improves insulin sensitivity, which is a large part of why these drugs move A1C at all. If you want the weight side of the picture, the GLP-1 weight loss calculator runs the same kind of estimate on pounds.
Limitations
This is arithmetic on a published regression line, not a lab. Your real A1C can land a few tenths either side of what the equation predicts from your meter average, and the conditions listed above can push it further. It estimates. It does not diagnose, and it can’t tell you what your A1C will do on a medication, only your next lab draw does that. Take the result to the person who ordered the test.