The best flours for blood sugar are almond meal, lupin, soy and coconut flour, because they have the least carbohydrate: 5 to 26 g per 100 g, against 73 g in plain white flour. Next come chickpea flour (39 g) and rye flour (55 g, with 17 g of fibre). White wheat flour, rice flour and gluten-free plain flour sit at the bottom, at 72 to 81 g. Almost no flour has its own tested GI, because nobody eats flour raw. What we do have is the GI of breads and baking made from them, and there the lowest Australian results came from lower carb, soy, rye and seeded breads (24 to 53), while white and wholemeal wheat breads tested 64 to 78.
Below, 14 flours are ranked using the data that exists: carbohydrate and fibre per 100 g, and the GI of foods made with each one. After that, we explain why your own glucose response to a loaf can differ from the tables.
Ranked by available carbohydrate (the starch and sugar your body turns into glucose), lowest first, with fibre breaking near ties. Nutrition is from the FSANZ Australian Food Composition Database unless noted. GI values are from the University of Sydney GI database. Low GI is 55 or under, medium 56 to 69, high 70 or more.
| Flour | Carbs | Fibre | Protein | GI of foods made with it |
|---|---|---|---|---|
| 1. Almond mealProtein 19.7 g | 5.4 g | 11.0 g | 19.7 g | Not tested |
| 2. Lupin flourProtein 36.9 g (lupin flakes) | 10.0 g | 27.0 g | 36.9 g | Not tested as flour. Lupin bread gave about a third less glucose than white bread in one trial |
| 3. Soy flourProtein 49.8 g | 14.6 g | 16.0 g | 49.8 g | 40 soy flour bread (Australia, 2005) |
| 4. Coconut flourProtein 19.3 g (Niulife label) | 26.2 g | 38.5 g | 19.3 g | 32 to 68 wheat and coconut flour baking (Philippines, 2002) |
| 5. Chickpea flour (besan)Protein 19.7 g | 38.5 g | 20.8 g | 19.7 g | 55 to 68 chickpea flour breads (Australia, 2004) |
| 6. Rye flourProtein 11.4 g | 54.9 g | 17.3 g | 11.4 g | 48 to 53 most Australian rye breads |
| 7. Oat flourProtein 12.2 g (rolled oats) | 54.5 g | 9.5 g | 12.2 g | 54 oat flour muffin (Singapore, 2015) |
| 8. Barley flourProtein 9.9 g (pearl barley) | 60.6 g | 13.1 g | 9.9 g | 50 to 85 breads with barley flour (overseas) |
| 9. Spelt flourProtein 18.7 g | 59.1 g | 7.4 g | 18.7 g | 54 to 66 spelt breads |
| 10. Buckwheat flourProtein 12.8 g (groats) | 58.0 g | 5.0 g | 12.8 g | 47 to 72 buckwheat breads |
| 11. Wholemeal wheat flourProtein 11.6 g | 61.6 g | 8.8 g | 11.6 g | 64 to 78 Australian wholemeal breads |
| 12. Rice flourProtein 7.3 g | 71.6 g | 2.2 g | 7.3 g | Not tested on its own |
| 13. White wheat flour, plainProtein 10.0 g | 73.2 g | 2.2 g | 10.0 g | 70 to 74 Australian white breads |
| 14. Gluten-free plain flourProtein 3.4 g | 80.5 g | 1.7 g | 3.4 g | Varies widely by recipe |
Cornflour (84 g of carbohydrate per 100 g), tapioca (79 g) and arrowroot (79 g) are almost pure starch, with little or no fibre or protein. They're fine for thickening a sauce, but they're the worst base for baking if blood sugar is your concern.
A GI test needs people to eat a portion with 50 g of available carbohydrate, and nobody eats flour raw. So the University of Sydney tests foods as they are eaten. That's why the GI figures quoted for almond or coconut flour on blogs aren't in the international GI tables.
What the tests do show is how a flour changes the bread, cake or pancake it goes into. Four things matter:
As Dr Michelle Woolhouse, Vively's Medical Director, puts it: "A diet rich in fibre from vegetables, fruits, whole grains, berries, nuts and legumes helps regulate blood sugar levels."
The lowest carbohydrate flour, but it's mostly fat (50.5 g per 100 g), so it has 2,386 kJ per 100 g against 1,479 kJ for plain wheat flour. It suits cakes, muffins, biscuits and crumb coatings, where you don't need a yeast rise, and recipes usually add extra egg to bind it. There's no published GI. Leave it out if you have a tree nut allergy.
Lupin flour is milled from the kernel of sweet lupin, a legume grown mostly in Western Australia. It's very high in protein (about 37 g per 100 g) and fibre. In a University of South Australia trial (Keogh 2011), the glucose response to lupin bread was about a third lower than to white bread, and people felt fuller. In an earlier Deakin University trial (Hall 2005), adding lupin flour lowered the GI of a white bread breakfast from 100 to 74 on the white bread scale, but raised the insulin response. Lupin is related to peanut, and some people with a peanut allergy react to it, which is why it has had to be declared as an allergen on Australian labels since 2017.
The highest protein flour (49.8 g per 100 g). An Australian soy flour bread tested GI 40, and loaves that swap some wheat for soy, such as Burgen Soy-Lin (36) and Helga's Lower Carb Soy and Toasted Sesame (51), also tested low.
Coconut flour is the highest fibre flour on the list, using the Niulife label. It soaks up far more liquid than wheat flour, so start with a recipe written for it. A 2002 Philippine study tested baked goods made with a mix of wheat and coconut flour at GI 32 to 68, from macaroons at the low end to a sweet bread roll at the top.
Good for pancakes, fritters, flatbreads and pakoras, with 20.8 g of fibre and 19.7 g of protein per 100 g. In an Australian study, chickpea flour bread tested 55 when made from desi chickpea flour and 68 from extruded chickpea flour. A white bread with only 20% chickpea flour still tested 80, so a small amount doesn't fix a white loaf.
Rye flour has twice the fibre of wholemeal wheat flour. Rye sourdough tested 48, Burgen Rye 51 and Country Life organic rye 53. Not every rye loaf is low: one dense Sydney rye tested 86 in 1989.
Because the GI attaches to the loaf, the most useful thing to compare is the bread on the shelf. Filter the list below by GI band, sort it, or pick the bread you buy now to see a lower GI swap.
None of these flours contains gluten, the protein that lets wheat dough stretch and trap gas. Swap all the wheat flour in a yeast bread for almond, coconut or lupin flour and you get a dense, crumbly loaf that won't rise. For bread, bakers usually replace part of the wheat flour with lupin, chickpea or soy flour, which keeps the structure and still cuts the starch. If you want bread for toast and sandwiches, buying a tested low GI loaf is simpler. Our list of low GI breads in Australia names the ones with published values, and we compare grains in barley vs wheat flour and oats vs wheat.
A GI test feeds about 10 healthy people a fixed amount of carbohydrate from one food, after an overnight fast, and averages the results. You eat bread with butter, eggs or cheese, in your own portion. And people respond differently to the same slice, depending on sleep, stress, activity and metabolism. The GI is an average, and you're not an average.
Our members' continuous glucose monitor (CGM) data shows why. The GI tables put white bread and traditional sourdough far apart. In everyday use the gap between loaves was smaller than the gap between people.
The typical rise sat between 1.7 and 2.1 mmol/L. Sourdough was lowest and Turkish bread highest, but the middle half of members overlapped for every bread, so the gap between loaves is smaller than the gap between people.
| Value | Members | |
|---|---|---|
| Median rise (mmol/L): Turkish bread | 2.1 | 132 |
| Median rise (mmol/L): Croissant | 1.9 | 264 |
| Median rise (mmol/L): White bread | 1.8 | 151 |
| Median rise (mmol/L): Banana bread | 1.8 | 521 |
| Median rise (mmol/L): Sourdough | 1.7 | 522 |
For sourdough, the middle half rose between 1.0 and 2.6 mmol/L, and 27% rose less than 1 mmol/L. What you eat with the bread also counts, for some people more than others. Among 243 members who logged sourdough both on its own and with only protein, fat or vegetables alongside (eggs, avocado, cheese, salad), the typical rise was 1.7 mmol/L alone and 1.5 mmol/L with the extras, and 55% had a lower rise with them.
This is real-world app data, not a lab test. Portions weren't measured, and someone may have eaten something they didn't log. The members don't have type 2 diabetes. See each food's full curve on our pages for sourdough, white bread and Turkish bread.
No single flour is best for type 2 diabetes. What helps is less refined starch per meal and more fibre, protein and intact grain. In practice that means choosing a tested low GI bread, using part lupin, chickpea or soy flour in home baking, and keeping portions of any baked goods moderate.
Michelle's reminder: "Remember, we're in no way demonising carbohydrates, they are an essential macronutrient!" Glucose after a meal is also only half the story. Your body can keep glucose in range for years by making more insulin, and a fasting insulin test shows that extra effort before glucose itself looks abnormal. If you have diabetes, talk to your GP or dietitian before you change your carbohydrate intake, especially if you take insulin or medicines that can cause lows.
Published GI values tell you how a bread behaves on average. With a Vively CGM, you can see how your toast or your home-baked loaf behaves in you, and compare one bread with another in the app. Our guide to blood sugar after eating explains what the numbers mean.
Wear a Vively CGM and compare your own curves after the breads and baking you eat.
No flour has a tested GI of its own. The lowest in carbohydrate are almond meal (5.4 g per 100 g), lupin (10 g) and soy flour (14.6 g). Among breads, lower carb, soy, rye and seeded loaves tested lowest (24 to 53), led by Herman Brot Lower Carb Bread at 24.
Almond flour has no published GI, and the "GI under 1" seen online isn't from GI testing. With only 5.4 g of carbohydrate per 100 g, almond baking raises glucose less than wheat baking. It's high in energy, at 2,386 kJ per 100 g.
Coconut flour hasn't been GI tested on its own. Baked goods made from a mix of wheat and coconut flour tested 32 to 68 in a Philippine study. It has 26 g of carbohydrate and 38.5 g of fibre per 100 g (Niulife label).
For blood sugar they're similar, as both are low in starch. Lupin flour has more protein and fibre and much less fat, so it has fewer kilojoules. Almond flour bakes more easily in sweet recipes. Lupin can affect people with a peanut allergy, and almond is a tree nut.
Lupin flour is milled from sweet lupin kernels, a legume grown mainly in Western Australia. With about 10 g of available carbohydrate per 100 g, it suits low carb and keto baking better than any grain flour.
Barley flour itself isn't tested. Breads made with barley flour tested 50 to 85 depending on the recipe, while breads made with whole barley kernels tested much lower (28 to 44, Canada, 1988).
Atta is finely milled wholemeal wheat flour. Rotis made from it tested 45 in India and 54 to 68 in the UK. Atta mixes with added chickpea (Bengal gram) flour, psyllium and fenugreek tested as low as 27.
There isn't one best flour. Lower carb flours (lupin, almond, soy, chickpea) and high fibre rye help, as does buying a tested low GI bread. Talk to your dietitian about portions.
START YOUR BASELINE
Vively includes a Baseline Health Check with 70+ biomarkers, a results report and practitioner support in the $99 a month membership. No minimum term.
Explore Baseline Health CheckSubscribe to our newsletter & join a community of 30,000+ Aussies
Start a PCOS assessment with your GP. Learn which blood tests matter, when ultrasound or AMH is useful, and how contraception affects results.
Know what to do about an unexpected high blood sugar reading, when to seek urgent help, and which everyday habits support longer-term glucose care.
.avif)
.svg)
Vively is a general wellness platform and does not provide medical advice, diagnosis, or treatment. Any information, insights or guidance offered through our app, website, programs, or coaches is for educational and informational purposes only and should not be considered a substitute for professional medical advice or care. Always consult a qualified healthcare provider before making any health-related decisions.
We acknowledge the Traditional Custodians of the lands where we live and work, the Gadigal of the Eora Nation. We pay our respects to their Elders past and present.
Share
Tweet
Share
