Berberine, cinnamon and fenugreek all lower blood sugar modestly. Berberine does it as a drug.
Several plant preparations lower blood glucose modestly. The effect sizes are usually small — a reduction in fasting glucose and HbA1c of less than a typical drug's worth — and the trials are commonly small and of short duration. The framing that matters: these are adjuncts to a treatment plan, and diabetes is a disease where the complications are prevented by glycaemic control, blood pressure control, lipid control and screening for eye, kidney and foot disease. A herb that reduces HbA1c by 0.3% does not replace any of that.
Berberine has the strongest evidence of the group, with multiple small trials showing reductions in fasting glucose, postprandial glucose and HbA1c comparable in some trials to metformin. It works partly by activating AMPK. Two important cautions: it is a potent CYP3A4 and P-glycoprotein inhibitor, which raises the levels of many co-administered drugs, and it is not for use in pregnancy or breastfeeding (it is used traditionally as an abortifacient and it can displace bilirubin in newborns).
Cinnamon — pooled analyses show a small reduction in fasting glucose, with notable heterogeneity and a pattern where the larger and better-designed trials find less effect. Cassia cinnamon, the cheap common type, contains coumarin, which is hepatotoxic at high doses; Ceylon cinnamon does not. Anyone taking cinnamon daily for months should use Ceylon.
Fenugreek — seeds have small trials showing glycaemic benefit. It also has an obvious culinary use. It can cause GI upset and it smells strongly of maple.
Alpha-lipoic acid — trials in diabetic neuropathy (see that page) and some evidence for glycaemic effects.
Chromium, gymnema, bitter melon — small trials, inconsistent results, and bitter melon in particular can cause hypoglycaemia when combined with glucose-lowering drugs.
Magnesium and vitamin D — both have been associated with glycaemic measures, and correcting a deficiency is reasonable. Supplementing beyond sufficiency has not shown consistent benefit.
Inositol — has trials in polycystic ovary syndrome, where it improves insulin sensitivity and ovulation.
The intervention with the best evidence of anything on this page: weight loss (substantial weight loss can put type 2 diabetes into remission), diet quality, exercise, and taking the prescribed medicines. Hypoglycaemia is the main risk of adding any of these to a sulfonylurea or insulin — monitor more, not less, when adding a supplement.