The glycemic index (GI) concept was born in 1981 at the University of Toronto as a tool for diabetics. Today everyone uses it โ from people losing weight to athletes. But 80% of people apply GI wrong: they look only at the index and forget about portion size. Let's break down how GI actually works, what glycemic load (GL) is, and why watermelon at GI 75 is less "dangerous" than a baguette at GI 72.
The glycemic index is a number from 0 to 100 that shows how quickly a food's carbs raise blood glucose, relative to pure glucose (GI 100) or white bread (depending on the scale). The concept was developed by Professor David Jenkins at the University of Toronto to improve blood sugar control in diabetics.
The GI scale:
The idea is simple: low-GI foods = longer satiety + fewer insulin spikes + lower risk of type 2 diabetes. High-GI foods deliver fast energy (great for athletes after a workout), but carry a high risk of overeating and metabolic problems with regular consumption.
๐ก Key fact: A meta-analysis from the University of Sydney (Livesey et al., 2019) covering data from over 500,000 people found that lowering a diet's average GI by 10 points was associated with a 12% reduction in type 2 diabetes risk and a 7% reduction in cardiovascular mortality. GI works โ but only if you apply it correctly.
The main scientific gripe with GI: it's measured at a fixed 50 g serving of carbs. But in real life, you don't eat 50 g of carbs from watermelon (that's 750 g of flesh) โ you eat one slice, about 150 g, containing 12 g of carbs.
To fix this, the concept of glycemic load (GL) was introduced in 1997:
GL = (GI ร carbs in the serving in grams) / 100
The GL scale:
The classic example โ watermelon vs. baguette:
| Food | GI | Serving | Carbs | GL |
|---|---|---|---|---|
| Watermelon | 75 (high) | 150 g (slice) | 11 g | 8 (low) |
| Baguette | 72 (high) | 100 g | 52 g | 37 (very high) |
Conclusion: watermelon, with its scary GI of 75, creates 4.5 times less load than a baguette with GI 72. That's the whole point โ looking only at GI is incorrect.
Data from the international Harvard T.H. Chan School database. Servings are typical real-world portions:
| Food | GI | Serving | GL |
|---|---|---|---|
| White bread | 75 | 30 g (1 slice) | 11 |
| Whole-grain bread | 54 | 30 g | 7 |
| Baguette | 72 | 100 g | 37 |
| Oatmeal (rolled) | 55 | 250 g cooked | 13 |
| Instant oatmeal | 79 | 250 g cooked | 21 |
| Sweetened muesli | 66 | 30 g | 16 |
| White rice | 73 | 150 g | 29 |
| Basmati rice | 52 | 150 g | 18 |
| Brown rice | 55 | 150 g | 18 |
| Buckwheat | 54 | 150 g | 15 |
| Quinoa | 53 | 150 g | 13 |
| Durum-wheat pasta (al dente) | 44 | 180 g | 20 |
| Food | GI | Serving | GL |
|---|---|---|---|
| Apple | 36 | 1 medium (150 g) | 5 |
| Pear | 38 | 1 medium | 4 |
| Orange | 43 | 1 medium | 5 |
| Banana (yellow) | 51 | 1 medium | 13 |
| Banana (with brown spots) | 62 | 1 medium | 16 |
| Watermelon | 75 | 150 g | 8 |
| Cantaloupe | 65 | 150 g | 7 |
| Pineapple | 66 | 150 g | 11 |
| Grapes | 56 | 120 g | 11 |
| Carrots, raw | 16 | 100 g | 1 |
| Carrots, boiled | 49 | 100 g | 3 |
| Potato, boiled | 78 | 150 g | 21 |
| Potato, baked | 85 | 150 g | 25 |
| Sweet potato | 63 | 150 g | 17 |
| Pumpkin | 75 | 80 g | 3 |
| Food | GI | Serving | GL |
|---|---|---|---|
| Kidney beans | 24 | 150 g | 7 |
| Lentils | 32 | 150 g | 5 |
| Chickpeas | 28 | 150 g | 8 |
| Whole milk | 31 | 250 ml | 4 |
| Plain yogurt | 14 | 200 g | 1 |
| Ice cream | 57 | 50 g | 6 |
| Food | GI | Serving | GL |
|---|---|---|---|
| Cola | 63 | 250 ml | 17 |
| Milk chocolate | 43 | 50 g | 12 |
| Dark chocolate (70%+) | 22 | 50 g | 4 |
| Honey | 61 | 25 g (1 tbsp) | 12 |
| Table sugar | 65 | 10 g (2 tsp) | 6 |
| Potato chips | 54 | 30 g | 8 |
| Popcorn | 65 | 20 g | 7 |
This is a popular diet built on a false premise. Yes, low-GI foods keep you full longer, which helps you eat fewer calories. But the law of thermodynamics is still in effect: if you eat more calories than you burn, you'll gain weight regardless of GI.
A 2007 Cochrane Review (Thomas et al.) of 6 studies with 202 participants found that low-GI diets were more effective for weight loss by 1 kg over 8โ26 weeks. There's an effect, but it's modest. The main thing in any diet is the calorie deficit; GI is a supporting tool.
Watermelon and pumpkin with high GI are not a problem because their GL is low. Rice and potatoes with high GI and high GL are a real reason to limit portions. That's the core principle: GI ร portion = real-world effect.
The same foods have different GIs depending on how they're prepared:
In practice: chill cooked rice or potatoes before eating โ GI drops 15โ20%.
Protein and fat slow carbohydrate absorption. Adding 10 g of fat or 15 g of protein to a meal cuts the peak glucose spike by 20โ40% (Anderson et al., AJCN 2015).
Examples:
For athletes: high GI is helpful after training (rapid glycogen replenishment). Before training and on regular days, low GI provides steadier energy.
If you have type 2 diabetes or prediabetes, GL is one of the best tools for glucose control. The American Diabetes Association recommends keeping total daily GL below 100. That's achievable when you:
โ ๏ธ Important: GI doesn't account for individual glucose response. The same food can produce a 2โ3ร different reaction in different people (Weizmann Institute study, Cell 2015). Continuous glucose monitoring (CGM) is the most accurate way to understand your own responses, if it really matters for you (diabetes, prediabetes, metabolic syndrome).
Situations where the GI concept loses meaning:
๐ก Short version: GI is a useful tool, but only paired with GL and seen in the context of the whole meal. For a healthy person, focusing on food quality (whole foods, protein, fat, fiber) works better than memorizing GI charts. For diabetes and prediabetes, GL becomes critical.
Calculating GL by hand for every meal is slow and tedious, especially for mixed plates. The NutriAI Pro AI nutritionist identifies the GI/GL of every ingredient in a photo and shows the meal's total load. The first 2 analyses are free in Telegram.
๐ Try it in Telegram