GI was developed by David Jenkins and colleagues at the University of Toronto in 1981 to compare carbohydrate-containing foods by their blood glucose response. The methodology: feed subjects a fixed 50 g carbohydrate portion of a test food, measure blood glucose response over 2 hours, compare against the same subject's response to 50 g glucose.
The classification:
| GI value | Category | Examples |
|---|---|---|
| 70+ | High | White bread, white rice, instant oatmeal, watermelon |
| 56-69 | Medium | Whole wheat bread, brown rice, bananas, sweet potato |
| 0-55 | Low | Most legumes, most fruits, oats, dairy, most non-starchy vegetables |
What GI gets right:
- Standardized comparison: GI provides a consistent, reproducible scale for ranking foods.
- Diabetic context: GI is one input for managing blood glucose in diabetic populations.
- Athletic timing: pre- and post-workout food choices can be informed by GI for fueling and recovery.
What GI gets wrong (or is misused for):
- Fixed 50 g portion is not realistic: a 50 g carb portion of watermelon is nearly 4 cups, far more than anyone eats. The high GI of watermelon (around 76) is meaningless at typical 1-cup servings, where the 11 g of carbs produces a small glycemic response.
- GI of mixed meals is not the GI of individual foods: protein, fat, and fiber in a meal slow digestion and lower the meal's overall glycemic response below what the carb's individual GI predicts.
- Test variability is high: the same food can score 10-20 GI points different across studies due to ripeness, processing, cooking method, and individual variation.
- GI does not predict weight loss outcomes in non-diabetic populations once calories and protein are matched (multiple meta-analyses).
Glycemic load fixes most of GI's practical issues by multiplying GI by the actual carb amount in a serving. A high-GI food in a small portion (watermelon) ends up with low GL; a moderate-GI food in a large portion (white rice) ends up with high GL. See the [Glycemic Load](/glossary/glycemic-load) entry for the corrected metric.
Practical implications:
- For non-diabetics, do not chase low-GI foods. Calorie balance, protein adequacy, and fiber intake matter more for weight loss outcomes.
- For diabetics, GI is one input but full management considers GI, GL, total carbs, fiber, fat, protein, activity, and medications. Work with a registered dietitian.
- For athletes, higher-GI foods around training (pre, intra, post) can support fueling and glycogen restoration. Lower-GI foods at other meals support steady energy and satiety.