Written by: Dr. Albana Greca, MD, MMedSc, Family Physician and Medical Author.
Medically reviewed by: Dr. Ruden Cakoni, MD, Endocrinologist.
Last reviewed: July 2026.
Quick answer: The glycemic index (GI) compares how equal amounts of available carbohydrate from different foods affect blood glucose under standardized testing. Lower-GI foods generally produce a slower, smaller average response than higher-GI foods, but GI does not measure portion size, overall nutritional quality, or your exact personal response. Use it together with total carbohydrate, glycemic load, meal composition, and glucose monitoring.
The glycemic index can help explain why equal amounts of carbohydrate from different foods do not always produce the same glucose response. Lentils, firm pasta, and a whole apple are generally digested and absorbed more gradually than cornflakes, instant oats, or many varieties of potato and rice.
However, a GI number is not a command to eat or avoid a food. Values differ by variety, brand, processing, cooking, ripeness, and laboratory testing. The amount eaten often matters as much as—or more than—the GI category. This page explains how to interpret a glycemic index chart safely and use it as one part of practical diabetes meal planning.
GI is a relative ranking of foods that contain digestible carbohydrate. Under standardized testing, participants consume a portion of the test food providing a fixed amount of available carbohydrate. Their glucose response is compared with their response to a reference—usually glucose, assigned a GI of 100.
The published value represents the average relative response across participants. It does not mean that eating a GI-70 food will raise your meter by 70 mg/dL, and it does not predict the exact glucose peak you will experience.
GI is often described as the “speed” at which a food raises glucose. That is a useful simplification, but technically the calculation is based on the incremental area under the blood-glucose response curve over the testing period—not only the time of the peak.
GI values are usually grouped into three categories: low GI is 55 or less, medium GI is 56–69, and high GI is 70 or more. These categories describe the average relative glucose response to a standardized amount of available carbohydrate. They do not show how much carbohydrate is in your normal serving.
A food must contain enough available carbohydrate to be tested with the standard method. That is why reliable GI databases generally do not assign a GI to:
These foods may still affect the glucose response of a mixed meal through protein, fat, fiber, gastric emptying, calories, and portion size. “GI not applicable” does not mean “eat without limits” or “automatically healthy.” The broader nutritional quality of a food still matters when building a practical diabetes meal plan.
The values below are representative tested values or common tested ranges, not universal numbers for every brand, variety, recipe, or country. When a product has been formally tested, use the value for that specific product. Otherwise, use the table as an educational comparison alongside our guide to low-glycemic-index foods.
| Food | Representative GI | Category | Why the value may differ |
|---|---|---|---|
| Natural yogurt | 10–19 | Low | Sweetened products and thickeners can change the result |
| Lentils, boiled | About 16–30 | Low | Canned lentils may test somewhat higher |
| Soybeans, boiled | About 20 | Low | Preparation and product type matter |
| Milk | About 20–34 | Low | Flavored milk may differ |
| Pearl barley, cooked | About 25 | Low | Processing and cooking time affect starch availability |
| Split peas, cooked | About 25 | Low | Whole versus puréed preparation may differ |
| Cannellini beans, canned | About 31 | Low | Brand, sauce, and preparation vary |
| Chickpeas, canned and drained | About 35–40 | Low | Home-cooked and canned products may differ |
| Red kidney beans, canned | About 36–40 | Low | Sauce and processing affect the value |
| Apple, whole | About 40 | Low | Variety, ripeness, and processing matter |
| Strawberries | About 40 | Low | Many other berries contain too little carbohydrate for reliable testing |
| Fresh rice noodles, cooked | About 40 | Low | Rice type and noodle processing vary |
| Steel-cut oats | About 42–52 | Low | Brand and cooking duration matter |
| Soba or buckwheat noodles | About 46 | Low | Check whether the product contains mostly buckwheat or refined wheat |
| White spaghetti, boiled | About 47 | Low | Overcooking and portion size raise the meal’s impact |
| Baked beans | Often about 48–49 | Low | Recipe and sauce can produce substantial variation |
| Watermelon | About 50 in the 2021 tables | Low | Older tests reported higher values; varieties differ and normal servings have a low GL |
| Mango | About 51 | Low | Ripeness and variety affect the result |
| Banana | About 52 | Low | Riper fruit may produce a higher response than less-ripe fruit |
| Quinoa, cooked | About 53 | Low | Product and cooking method vary |
| Traditional sourdough bread | About 54 | Low | Commercial “sourdough-style” products may not test the same |
| Papaya | About 56 | Medium | Variety and ripeness matter |
| Traditional rolled oats | About 57–58 | Medium | More intact oats generally test lower than instant products |
| Pineapple | About 59 | Medium | Fresh, canned, juice, and syrup-packed forms differ |
| Basmati rice | About 43–69 | Low to medium | Cultivar, amylose content, cooking, and origin create real variation |
| Cantaloupe or rockmelon | About 68 | Medium | A normal serving may still have a low glycemic load |
| Arborio or risotto rice | About 69 | Medium | Preparation and portion size influence the meal response |
| Standard white bread | About 70 | High | Dense, grainy, fermented, and legume-enriched breads may test lower |
| Boiled potato | Average about 73 | High | Variety, cooking, cooling, and mashing can change the response substantially |
| Cornflakes | Often about 77–93 | High | Brand formulation and processing differ |
| Baked potato | About 80 in a common example | High | Potato variety creates a wide range of tested values |
| Instant oats | About 82 | High | Smaller particle size and processing speed digestion |
| Jasmine rice | About 68–109 | Medium to high | Cultivar, ethnicity of participants, cooking, and testing conditions vary |
| Rice crackers | About 91 | High | Puffing and processing make starch readily digestible |
| Water, plain meat, fish, eggs, oils, most cheese, avocado, and leafy vegetables | Not applicable | Not testable by standard GI method | Contain little or no available carbohydrate |
Chart caution: The chart is for comparison, not insulin dosing. Verify the brand or food in the University of Sydney GI database when a tested value is important, and monitor your own response when your clinician recommends it.
Rice and potatoes show some of the largest variation. Foods with more amylose generally digest more slowly than foods with more amylopectin. Basmati rice often tests lower than jasmine or glutinous rice, but not every basmati product is low GI. This is why a single number should not replace a broader comparison of carbohydrate quality and glycemic index.
Grinding, flaking, puffing, instant preparation, and puréeing can make starch easier for digestive enzymes to reach. Steel-cut oats therefore often test lower than instant oats, even though both began as oats. Less-processed choices are often included in a well-planned low-glycemic diet.
Longer cooking can soften food structure and gelatinize starch. Pasta cooked firm is generally lower GI than very soft pasta. Potato values vary with variety, cooking, cooling, and reheating.
As some fruits ripen, starch is converted to sugars and digestibility changes. The effect varies by fruit and does not mean that ripe fruit must be avoided. Variety, serving size, and ripeness are also important when comparing lower-GI fruit choices.
Organic acids in traditional sourdough, yogurt, lemon juice, and vinegar may slow gastric emptying or change starch digestion. The size of the effect depends on the complete meal.
Two breads, yogurts, cereals, or rice products with similar names may contain different ingredients and have different physical structures. Color is not enough: brown bread is not automatically whole grain or low GI.
Blood glucose is affected by more than table sugar. Most digestible carbohydrate is broken down into glucose, including carbohydrate from bread, rice, pasta, potatoes, fruit, milk, and cereals. For this reason, meal planning should consider the total amount of carbohydrate, its source, fiber content, degree of processing, and the rest of the meal—not only whether a food tastes sweet.
Whole grains, legumes, whole fruit, and vegetables often contain fiber or retain more of their natural structure, which can slow digestion. Refined or highly processed carbohydrate may be digested more rapidly. Learn more in our guide to carbohydrates and diabetes.
GI does not increase when you eat a larger portion. The food’s GI remains the same, but the total glucose response usually increases because you consumed more carbohydrate. This distinction is central to understanding the relationship between carbohydrate intake and diabetes.
Glycemic load combines GI with the available carbohydrate in the portion:
GL = GI × available carbohydrate in the serving ÷ 100
An apple with a GI near 40 and about 15 grams of carbohydrate has a GL of approximately 6. A small baked potato with a GI near 80 and 15 grams of carbohydrate has a GL of approximately 12.
Use our glycemic load chart and calculator to compare actual portions.
Low-GI or low-GL dietary patterns may provide a modest additional improvement in glycemic control when compared with higher-GI or control diets. A 2021 systematic review and meta-analysis of randomized trials reported an average HbA1c reduction of approximately 0.31 percentage points. A practical starting point is to replace highly processed carbohydrates with suitable foods from a lower-GI food list.
This does not mean GI is a treatment by itself. The benefit depends on the complete eating pattern, portion size, adherence, medicines, activity, baseline HbA1c, and overall food quality.
The American Diabetes Association describes GI and GL as tools that may add information beyond total carbohydrate. ADA also emphasizes individualized nutrition rather than one required percentage of carbohydrate, protein, and fat for every person.
Foods that often support a steadier post-meal response include non-starchy vegetables, beans, lentils, chickpeas, barley, less-processed oats, whole fruit, nuts, seeds, unsweetened yogurt, fish, eggs, and other suitable protein foods. Foods without meaningful carbohydrate do not have a useful GI value, but they can still affect calories, fullness, gastric emptying, and the timing of a mixed-meal response. For more examples, see our guides to low-glycemic-index foods and lower-GI fruits.
For practical meal guidance, see our low glycemic diet guide and diabetes food and meal-planning guide.
Protein and fat do not change the published GI of the carbohydrate food. They can change the glucose response of the meal by affecting gastric emptying, insulin secretion, carbohydrate amount, and digestion. For this reason, results should be interpreted together with your usual blood sugar levels after eating.
This effect is not always protective. A high-fat mixed meal may produce a smaller early rise followed by delayed, prolonged hyperglycemia. This is particularly important for people using mealtime insulin or living with gastroparesis.
Adding butter or oil should not be used as a strategy to make a high-GI food “healthy.” Favor unsaturated fats, appropriate portions, vegetables, and nutrient-dense foods.
No. Some chocolate, ice cream, fried food, and high-fat snack products may have a lower GI because fat slows gastric emptying. They may still be high in calories, saturated fat, sodium, or added sugar.
GI also cannot tell you whether a food provides protein, fiber, vitamins, minerals, or an appropriate amount of sodium and saturated fat. Evaluate the whole food and the complete meal.
No. A higher-GI food can fit in a suitable portion, particularly when the meal includes vegetables, protein, and lower-GI carbohydrate choices.
Some high-GI carbohydrates are useful for treating hypoglycemia, but use a measured fast-acting source from your low-glucose plan—such as glucose tablets, glucose gel, or regular juice. Do not choose treatment only by consulting a GI chart; chocolate, pastries, and high-fat foods may act too slowly.
GI charts are based on average responses measured under standardized conditions, so they are useful for comparison but cannot predict every individual result. Your response may differ because of diabetes type, insulin sensitivity, medicines, activity, sleep, stress, illness, time of day, portion size, food combinations, and digestive conditions.
Regional varieties and traditional recipes may also differ from foods tested in published databases. When a specific branded product has been formally tested, use that product’s result. Otherwise, treat chart values as representative estimates and confirm important patterns with your meter or CGM.
Two people can respond differently to the same food, and one person can respond differently on separate days. Looking at repeated readings rather than one isolated result can help you understand whether a meal produces a consistent post-meal glucose pattern. Factors include:
To compare foods, keep portions and timing reasonably consistent, record the meal, and repeat the observation before drawing a conclusion. Use our blood sugar log to identify repeated patterns.
Changing from a usual carbohydrate pattern to much lower-GI or lower-carbohydrate meals may reduce glucose exposure and alter medication needs. Review any major dietary change with your diabetes team, particularly when following a low-glycemic eating plan. Do not reduce insulin, sulfonylureas, or meglitinides without a written plan or clinician guidance.
Pregnancy uses tighter glucose targets. A low-GI approach may be one component of nutrition care, but it does not replace pregnancy-specific carbohydrate distribution, glucose monitoring, and obstetric guidance.
Some low-GI foods are high in potassium or phosphorus. Kidney nutrition should be individualized rather than based on GI alone.
Large amounts of fiber and fat may delay stomach emptying and worsen symptoms. A standard low-GI plan may need modification.
Do not calculate insulin from a GI value. Insulin dosing is based on prescribed carbohydrate ratios, correction factors, glucose trends, active insulin, and other individualized instructions.
Treat low glucose promptly according to your written plan. Confusion, seizure, unconsciousness, inability to swallow safely, vomiting with ketones, deep breathing, or severe dehydration requires urgent or emergency care.
I use GI as a comparison tool, not a food-grading system. It is most useful when comparing two realistic carbohydrate choices—such as instant versus steel-cut oats or jasmine versus a tested lower-GI rice. I then consider the portion, the rest of the meal, nutritional quality, medicines, and the patient’s actual glucose response.
The glycemic index helps explain why some carbohydrate foods produce a faster average glucose response than others. It works best as a comparison tool—not a food-grading system. Combine GI with portion size, total carbohydrate, glycemic load, food quality, meal composition, medicines, activity, and your own repeated glucose readings.
There is no single daily GI target required for everyone. Choosing lower-GI carbohydrate foods more often can help, but portions, total carbohydrate, nutritional quality, and personal glucose patterns also matter.
Yes. A GI of 55 or less is classified as low.
No. Tested brown-rice values range from low to high depending on cultivar, amylose content, cooking, origin, and testing population.
It is more accurate to say GI is not applicable. Plain meat contains too little carbohydrate to be tested with the standard GI method.
No meaningful GI is assigned to eggs because they contain very little carbohydrate.
Older tests reported high values, but the 2021 International Tables report an average GI of about 50 across four tested varieties. A normal serving also has a low glycemic load.
It can. Longer cooking or greater processing may make starch easier to digest, but the effect depends on the food.
No. GI is not grams of carbohydrate and should not be entered into an insulin calculator.
Both can matter, but carbohydrate amount is essential for meal planning and insulin dosing. GI provides additional information about carbohydrate quality.
Educational safety note: This article provides general diabetes nutrition education. It does not replace personal medical nutrition therapy, diagnosis, glucose-monitoring instructions, or treatment. Do not start, stop, or change insulin, diabetes medicines, supplements, diet, or exercise based only on a GI chart.