Gi Calculator

Most people who search for a GI calculator want one of two things: the glycemic index of a food they are about to eat, or the glycemic load of a real serving of it. This page does both. Pick from 56 common foods whose GI values come from the international tables, adjust the servings, and the tool returns the load per serving and in total. Or type in any GI and carbohydrate figure of your own for a food that is not listed.

Quick Answer

How to use the GI calculator

Three entries drive the result: a food, its carbohydrate per serving, and how many servings you eat. The lookup supplies the first two for 56 foods, and the custom mode covers everything else.

1. Pick a food, or go custom. Selecting a food fills the GI and carbohydrate boxes from the lookup table and shows its typical serving. To run a food that is not listed, select any food and overwrite the GI and carbohydrate boxes with your own figures from a reliable table or a GI-tested product label. That is the custom mode: the tool never locks you to the list.
2. Check the carbohydrate figure. GL uses available carbohydrate, meaning total carbohydrate minus fibre. That is the same figure the keto carb calculator calls net carbs. If your serving is bigger or smaller than the typical one shown, scale the carbohydrate grams to match before calculating.
3. Enter your servings. One is the typical serving. Enter 2 for a double portion or 0.5 for half. The headline result is the total load for that amount, because portion is the variable people most often get wrong.
4. Read both badges. The GI badge bands the food itself. The GL badge bands one serving, so you can see the food’s speed and the serving’s size effect side by side before portion multiplies anything.
5. Copy or share the result from the results box. The copied text restates the food, GI, carbohydrate and both loads, so the number never travels without its inputs.

What the glycemic index measures, and how a lab gets the number

The glycemic index ranks a carbohydrate food by the blood glucose rise it causes in the two hours after eating, scored against pure glucose at 100. It is a measured property of the food’s carbohydrate, not a calculation from its nutrition label.

The method follows the international standard the University of Sydney glycemic index group describes: ten or more healthy volunteers eat a portion of the test food containing 50 grams of available carbohydrate, blood samples are taken over the next two hours, and the area under each person’s glucose curve is divided by the area from an equal-carbohydrate glucose drink taken on another day. Average the volunteers and you have the food’s GI. Two consequences follow. First, GI compares foods at equal carbohydrate, never at equal servings, which is exactly the gap glycemic load was built to close. Second, GI is a human measurement with real spread: the same food tested in different labs returns a range, and published tables print the average.

The values here rest on the international tables: Atkinson and colleagues catalogued over 2,400 foods in the 2008 Diabetes Care edition, and the 2021 review in the American Journal of Clinical Nutrition expanded the list past 4,000 items. The Oregon State University Linus Pauling Institute review of glycemic index and load cross-checks the foods both sources list.

The glycemic load formula, and why GI on its own misleads

Glycemic load multiplies a food’s GI by the carbohydrate in one serving and divides by 100. That single step converts a laboratory ranking into a portion-aware number.

Watermelon is the classic teaching case, and it is the example pre-filled above. Its carbohydrate digests fast, so its GI is high at 76. But watermelon is mostly water: a 120 g serving carries about 6 g of available carbohydrate. Run the formula and the load is 76 × 6 ÷ 100 = 4.6, firmly low. A doughnut tests at the same GI of 76, yet a medium doughnut packs 23 g of carbohydrate, so its load is 17.5, at the top of medium. Same index, four times the load. Any page that ranks foods by GI alone puts those two side by side as equals, which is why this calculator leads with load and treats GI as the supporting figure.

Worked example: three foods, one serving each

One serving each of watermelon, a doughnut and white rice shows the whole GI versus GL argument in three rows. Watermelon and the doughnut share a GI; the rice has a lower GI than both and the highest load on the table.

Check each bar against the calculator. Watermelon is the pre-filled default: 76 × 6 ÷ 100 = 4.6. Select the doughnut and you get 76 × 23 ÷ 100 = 17.5. Select white rice and 66 × 53 ÷ 100 = 35.0, a high load driven by the 53 g of carbohydrate in a cup of cooked rice, not by speed: at 66 its GI is only medium. Portion is the lever in all three cases, which is why the servings box sits in the calculator rather than in a footnote.

GI and GL bands: where low, medium and high begin

The bands are the standard cut points used across the international tables and the Harvard and Sydney guides. GI bands describe the food; GL bands describe one serving of it.

The calculator wires these exact thresholds into its badges. One caution the bands cannot show: a low score is not a health verdict. Chocolate lands low because fat slows its digestion, and peanuts sit near the bottom of the scale while carrying plenty of calories. The bands rank glucose effect only.

The food lookup table: GI, serving and carbohydrate for 56 foods

The same values the calculator uses, printed in full so the page works as a lookup even without calculating. GI values are the international-table averages; the GL column is computed from the GI and carbohydrate shown, at one typical serving.

Reading the table: fruit, legumes and dairy cluster low because their carbohydrate comes packaged with water, fibre or fat. Cereals, rice and potato run high because a normal serving is mostly starch. Baked russet potato has tested as high as 111 in one dataset, above glucose itself; the table uses 85, the value the Harvard chart prints. Boiled potato at 78 against instant mashed at 87 shows what processing does to the same vegetable.

Why the same food can test at two different GIs

A food does not own one fixed GI. Ripeness, cooking, processing and company on the plate all move the number, sometimes by 20 points or more.

Ripeness. A firm, just-yellow banana tests lower than a soft, spotted one, because ripening converts starch into sugars that absorb faster. The table value of 51 sits between the two states. Cooking time. Pasta boiled firm scores lower than the same pasta boiled soft: the Linus Pauling table puts spaghetti boiled 20 minutes at 58, against 46 across all cooking times, while this page uses the Atkinson et al. 2021 mean of 49. Processing. Milling, puffing and mashing pre-break the structure digestion would otherwise work through, which is why instant mashed potato (87) outruns boiled potato (78). Cooling. Cooked starch that cools forms resistant starch, so potato salad and reheated rice tend to test lower than the same food eaten hot. What it is eaten with. Fat, protein, fibre and acid slow gastric emptying: the Harvard Health guide to the glycemic index notes that olive oil or something acidic such as vinegar can blunt a food’s glucose rise. A table value is the food eaten alone, on an empty stomach, in a lab. Your plate is none of those three.

Adding up a mixed meal: sum the loads

A meal’s glycemic load is the sum of its parts. Work out each carbohydrate food separately, then add the loads; foods with no available carbohydrate add nothing.

Take a simple plate: one cup of cooked white rice (GL 35.0) with one cup of boiled lentils (GL 7.0). The meal load is 42.0. Swap the rice for one cup of boiled pearled barley (GL 10.6) and the same plate totals 17.6, less than half, with the lentils unchanged. That swap, rather than any single food ban, is how GL earns its keep in planning. The carbohydrate calculator helps when you first need the carbohydrate grams of each component, and readers working to a daily carb budget can set the meal in context with the diabetic carbs per day calculator. One honest limit: fat and protein slow the actual glucose rise, so a summed load is best used to compare meal designs, not to forecast a meter reading.

Who uses GI and GL, and where the medical guidance actually stands

GI and GL are used mainly in diabetes meal planning, and the guidance is more split than most calculator pages admit. Diabetes organisations in Australia, Canada and the UK include low-GI eating in their nutrition advice, and the British Dietetic Association glycaemic index guide presents GI as a practical tool for people with diabetes, pre-diabetes and gestational diabetes. The American Diabetes Association takes a cooler line: its Standards of Care does not recommend GI as a standalone method, and the ADA’s Diabetes Food Hub explains why the glycemic index is not a helpful standalone tool, pointing readers to carbohydrate counting and the plate method first, because individual responses vary and the original testing was not done in people with diabetes.

Used inside those limits, the practical pattern is simple: count carbohydrate first, then prefer the lower-GL version of similar foods where the swap costs you nothing: barley or lentils in place of part of the rice, firm pasta instead of soft, whole fruit instead of juice. The index ranks; the load portions; the rest of the meal still counts.

What competing glycemic calculators miss

The rival tools for this search are competent at one multiplication and stop there:

• Omni Calculator’s glycemic load tool computes GL correctly, but it starts where this page ends: you must already know the food’s GI and its carbohydrate per serving before the tool can help. There is no sourced food lookup to get those numbers from, so most visitors leave to find a table elsewhere and never come back to the calculator.
• OnlyCalculators’ glycemic index tool presents GI values without the serving arithmetic that makes them usable. A GI list on its own repeats the watermelon error at scale, ranking foods by speed while hiding the portion effect that decides the actual glucose load.
• The formula-only pages state GL = GI × carbs ÷ 100 and give one example. They skip the bands, skip the servings multiplier, skip custom entry for foods off their short lists, and skip the mixed-meal method entirely, so a two-food plate is beyond them.
• The gap this page fills: a 56-food lookup whose every GI and carbohydrate figure traces to the international tables, a servings multiplier that separates per-serving load from total load, band badges for both GI and GL wired to the standard thresholds, a custom mode for unlisted foods, and the honest ADA counterweight that most rival pages omit.

Limitations: what these numbers cannot tell you

GI and GL rank the glucose effect of carbohydrate foods. That is a narrow job, and the scores say nothing direct about how healthy a food is overall.

A low score can ride on fat: chocolate and ice cream test low to medium because fat slows digestion, which does not make them staple foods. A high score can ride on water: watermelon is the standing example. GI values come from small groups of healthy volunteers, and your own response shifts with sleep, activity, medication and the meal before, so two people can eat the same serving and record different curves. Values also move between brands: white bread alone has published tests from 59 to 89. Use the numbers to rank swaps between similar foods and to sanity-check portions, not as a score for a whole diet or a substitute for the monitoring your own care team recommends.

Frequently asked questions

Add This Calculator to Your Website

References