Stevia vs Erythritol: Sweetness and Bulk Are Two Different Jobs in Sugar Reduction
Stevia vs Erythritol is often presented as a choice between two sugar alternatives, but in commercial formulation they usually perform different jobs. Stevia restores sweetness at very low use levels, while erythritol restores part of the bulk and solids that disappear when sugar is removed.
A high intensity sweetener returns sweetness and almost no mass. A bulk polyol returns mass and only part of the sweetness. Sucrose delivered both at once, along with half a dozen physical functions that neither ingredient replaces on its own.
Choosing between them is the wrong decision. The decision that matters is which of sugar's jobs your product actually needs back, and which ingredient is doing each one.
One disclosure before anything else, because it should change how you read the rest of this page. We supply both steviol glycosides and erythritol, and we produce blended systems of the two. We have no commercial reason to steer you toward either ingredient, which is why several sections below say things that a single ingredient supplier would leave out.
This guide is not a health ranking. If you want to know which sweetener is better for a consumer, there are a thousand articles for that. What follows is the version a formulator and a purchasing manager need: what each ingredient physically does, where the two genuinely interact, what erythritol costs you in cooling and crystallization, what the dose data actually say per serving, where the cardiovascular literature stands in 2026, and how differently the two are regulated and priced in the EU and the US right now.
01 Framing
Stevia vs Erythritol: Two Ingredients, Two Different Jobs
Ask what happens if you remove one of them entirely.
Remove erythritol from a tabletop sweetener and the consumer receives a barely visible dusting of powder in a sachet designed to hold two grams. Remove erythritol from a sugar free cake and the batter has no solids to hold air, so it bakes into something dense and tight. Remove erythritol from a hard candy and there is nothing left to form the candy.
Remove the steviol glycoside from any of those and you get a product that is roughly the right size and shape but is only about two thirds as sweet as it should be, and would need so much more polyol to close the gap that it stops being a viable product for reasons covered in the section on dose and tolerance.
That is the whole argument in two paragraphs. The rest of this page is the detail behind it.
02 Arithmetic
The Mass Balance Nobody Runs Before the First Bench Trial
Take a beverage originally sweetened with 10 percent sucrose and ask what it takes to put the sweetness back.
| Component | Amount Needed to Match 10 Percent Sucrose | What It Also Contributes |
|---|---|---|
| Sucrose, baseline | 100 g/L | 100 g/L of dissolved solids |
| Reb A, about 250x | roughly 400 mg/L | Sweetness only |
| Reb M, about 300x | roughly 330 mg/L | Sweetness only |
| Erythritol alone, 0.65x | roughly 154 g/L | Solids, but 51 g in a 330 ml can |
| Erythritol at the EU beverage maximum | 16 g/L | About 1 percent sucrose equivalence, one tenth of the target |
The last two rows are the ones worth writing on a whiteboard.
Erythritol cannot carry full sugar replacement in a beverage on its own. At 154 g/L a single 330 ml serving delivers 51 grams, which sits at the top of the published tolerance range for a bolus dose and well above any sensible per serving target. In the European Union the question does not even arise, because the legal ceiling in flavoured drinks is 1.6 percent, which returns about a tenth of the sweetness you removed.
Steviol glycosides cannot carry the physical side of it at all. Four hundred milligrams per litre against one hundred grams per litre is a ratio of about 1 to 250. The sweetness comes back. The solids do not.
The pair is not a marketing convention. In beverages it is close to arithmetically forced.
How to read these figures
The sweetness multipliers used here are working approximations. Relative sweetness for any steviol glycoside shifts with concentration, matrix, pH and serving temperature, and published values for the same grade vary across sources. Use this table to size the problem, then measure sucrose equivalence in your own system before setting a dosage.
03 Diagnosis
Start With Sugar: What Was It Actually Doing in Your Formula?
Most reformulation projects begin with a percentage target handed down from marketing, and the formulation team is left to absorb whatever the target costs. A more productive starting point is a question about the original product.
Sucrose was doing between one and seven jobs in your formula depending on the category. Sweetness is only the first of them, and in several categories it is not even the most difficult one to replace.
| What Sugar Did | Steviol Glycosides | Erythritol | Where the Gap Usually Gets Closed |
|---|---|---|---|
| Sweetness intensity | Yes, fully | Partly, 0.6 to 0.7x | Blend of the two |
| Dissolved solids and bulk | No | Yes | Erythritol, polydextrose, soluble fibre, maltitol, isomalt |
| Viscosity and body | No | Limited | Soluble fibre, hydrocolloids, protein, glycerin |
| Water activity control | No | Yes, strongly | Erythritol, glycerin, sorbitol, fibre syrup |
| Freezing point depression | No | Yes, strongly | Erythritol, sugar retention, careful ice cream balancing |
| Browning and Maillard reaction | No | No | Sugar retention, reducing sugars, allulose where legal |
| Crystalline structure and bite | No | Yes, but recrystallizes | Isomalt, sugar retention, controlled seeding |
Five entries in the steviol glycoside column say no. That is not a weakness of stevia as an ingredient, it is a description of what a high intensity sweetener is. A material used at 300 parts per million cannot restore anything physical, no matter how good its taste profile is.
Using the table as a routing device
Work out which rows your product actually depends on, then look at which column can serve them.
- If sugar was mainly delivering sweetness. Flavoured waters, iced teas, low solids RTD drinks, some yoghurt drinks. Steviol glycosides can carry most of the replacement. You may need very little erythritol, and in the EU you may not be permitted to use much anyway. Put your effort into glycoside selection and the acid and flavour system instead.
- If sugar was delivering sweetness plus solids. Powdered beverages, tabletop sweeteners, protein powders, drink mixes, instant desserts. Stevia alone is not a product here. Something has to occupy the volume in the sachet or the scoop, and erythritol is the most sucrose like option available.
- If sugar was structural. Cakes, muffins, cookies, batters, chocolate, hard candy, gummies. Both ingredients together still leave you short, because sucrose in these systems also acts as a plasticizer, a humectant, a browning substrate and a control on starch gelatinization and water activity.
- If the product is prone to crystallization. Syrups, fillings, coatings, chewy confectionery. Erythritol dosage cannot be set from a sweetness calculation alone, because its solubility limit and storage temperature decide whether it stays in solution or turns the product sandy.
04 Stevia
What Steviol Glycosides Actually Contribute
Sweetness intensity at very low inclusion, and a sensory profile that depends heavily on which glycoside you bought.
The usage levels are what make the ingredient useful and what make it insufficient at the same time. A beverage takes a few hundred parts per million. A bakery formula takes a fraction of a gram where a kilogram of sugar used to sit. Nothing at that inclusion rate can contribute solids, viscosity, water activity control or browning.
The sensory side is not uniform across grades. Reb A, Reb D and Reb M are different molecules with different receptor behaviour, not different purities of one sweetener, and they should never be treated as interchangeable line items on a specification. Bitterness, sweetness onset and how long sweetness lingers all shift with the glycoside profile, and so does cost per unit of sweetness. Our stevia extract product page lists the grades we hold and the assay basis for each, which matters because 95 percent Reb A and 95 percent total steviol glycosides are entirely different materials at entirely different prices.
One number that surprises people who have only read consumer content: an independent receptor level comparison found the EC50 of Reb A at the human sweet receptor to be higher than that of Reb M, 56 micromolar against 34 micromolar, even though the two are usually quoted at similar sweetness potency. Receptor potency and perceived sweetness in a real matrix are related but not the same thing, and neither of them tells you anything about mass.
05 Erythritol
What Erythritol Actually Contributes
Bulk, plus about two thirds of sucrose's sweetness, plus a set of physical behaviours that come directly from its molecular size.
| Property | Value | Why It Matters in Formulation |
|---|---|---|
| Relative sweetness | 0.6 to 0.7x sucrose | You need more than a one for one weight swap |
| Molecular weight | 122, the lowest of the common polyols | Strongest colligative effects per gram of any polyol |
| Solubility in water | about 35 to 37 g per 100 g at room temperature | Roughly half that of sucrose, which sets the crystallization risk |
| Heat of solution | about minus 180 J/g, the largest negative value of any commercial polyol | The cooling sensation |
| Hygroscopicity | below 0.5 percent moisture uptake up to 90 percent RH | Excellent powder flow and caking resistance |
| Digestion | about 90 percent absorbed in the small intestine, excreted unchanged in urine, not fermented by gut microbiota | The reason it is the best tolerated polyol |
| Browning | none | Pale bakes, no crust colour development |
| EU energy value | 0 kcal/g, against 2.4 kcal/g for every other permitted polyol | A genuine and exclusive labelling advantage |
The molecular weight line is the one that explains most of the rest. At 122 daltons, erythritol produces more osmotic effect per gram than any other polyol in commercial use, which is why it depresses water activity strongly, depresses freezing point strongly, and why frozen and semi moist products behave differently than the sugar version even when the sweetness has been matched.
The cooling effect, and where it actually shows up
The cooling sensation is the single most common complaint and the most misunderstood. It is an endothermic dissolution effect, so it appears when erythritol crystals dissolve in the mouth and it largely disappears when erythritol is already dissolved before the consumer gets it. In a beverage you will barely notice it. In chocolate, frosting, powdered drink mix, chewable tablets, protein powder and hard candy it is unavoidable and needs to be either designed around or designed in. In peppermint gum and lozenges it is an asset that formulators pay for separately.
Three things reduce it in practice. Blending with ingredients that have a positive heat of solution, such as inulin, fructo oligosaccharides or glycerin. Reducing particle size, since finer crystals dissolve faster and shift the sensation. And simply using less erythritol, which is usually possible once the sweetness contribution is rebalanced toward the glycoside.
There is also a less known interaction in the other direction. Patent sensory work found that adding rebaudioside B to a 5 percent erythritol solution increased sweetness and reduced both bitterness and the perceived cooling effect. The relationship between these two ingredients is not one way.
06 Interaction
Why They Are Used Together, and Why Synergy Is Not Universal Masking
The synergy is real and it has been demonstrated at the receptor level, not only by tasting panels. Work carried out under the EU SWEET project used cells expressing the human TAS1R2/TAS1R3 sweet receptor to test six binary sweetener blends, and confirmed genuine synergy for the rebaudioside A and erythritol pair. The same study found no synergy at all for rebaudioside M combined with mogroside V, and proposed that those two compete for the same binding sites. Two natural sweeteners in the same bottle do not automatically reinforce each other.
The European Commission acknowledged the effect in legislation. The recitals to Commission Regulation (EU) 2015/1832, which authorised erythritol in flavoured drinks, state that at low levels erythritol acts as a flavour enhancer and helps mitigate the off tastes and lingering sweetness associated with high intensity sweeteners. When a regulator writes the mechanism into an authorising instrument, you can put it in a technical document without hedging.
Now the part that supplier content leaves out
A 2012 sensory study at Hamburg University of Applied Sciences optimised binary blends of erythritol with Reb A and with sucralose. At the optimised ratio, the blends produced sucrose like flavour and texture profiles. But when the researchers shifted those optimal combinations in order to push maximum sweetness higher, the characteristic unpleasant bitter notes of the sweeteners became perceptible again.
Synergy exists at a ratio, not as a property of the pair.
A blend that tastes clean at your pilot ratio can turn bitter when procurement asks you to raise the sweetness by fifteen percent without changing the cost. Two further limits are worth stating: that study used a 3 percent sucrose equivalent reference, which is a low sweetness model system and not a cola, and receptor level synergy is a sweetness effect, not a bitterness blocker.
So the sentence "add erythritol and the stevia aftertaste goes away" is not supported. What is supported is narrower and more useful. Erythritol raises the sweetness contribution of the glycoside at a given dose, which lets you use less glycoside, and using less glycoside is what actually reduces bitterness. The erythritol is helping indirectly, by allowing a lower stevia load.
If bitterness or lingering sweetness is the main limitation in your formula rather than body or bulk, the diagnosis and the fixes are a separate problem, and we covered them in detail in our guide to stevia aftertaste and glycoside selection.
07 Replacement level
What Changes Between 30 Percent and 100 Percent Replacement
Partial and total sugar replacement are not the same project with a different number attached. They fail in different places, and the published data show the transition clearly.
In water, the divergence point differs by sweetener
A 2023 temporal sensory study replaced sucrose at 25, 50, 75 and 100 percent in aqueous solution with acesulfame K, aspartame, erythritol, rebaudioside M and sucralose, and profiled each with a trained panel using temporal check all that apply. At 25 percent replacement, every sweetener except aspartame was indistinguishable from the sucrose control. Erythritol, aspartame and sucralose remained indistinguishable from sucrose at 50 and 75 percent replacement. Acesulfame K and rebaudioside M did not: above 25 percent replacement both produced significant differences in the aftertaste period, bitter for acesulfame K and licorice for rebaudioside M.
We sell rebaudioside M, and we are telling you that in a water model its temporal profile diverges from sucrose at a lower replacement level than erythritol's does.
That is what the data say. It is also a strong argument for using the two together rather than either alone, and for questioning whether your brief really requires 100 percent replacement.
In bakery, even the blend fails at full replacement
Two studies from Lincoln University tested a commercial stevia and erythritol blend in muffins at 50 and 100 percent sugar replacement. At 50 percent, texture was similar to the control and sensory acceptance was high. At 100 percent, the muffins were significantly firmer, more compact in the crumb, drier in the mouth and scored poorly with consumers.
Read that result carefully, because it is the most useful single data point on this page. The 100 percent samples did contain a bulking agent. Erythritol was in there doing its job. The formula still failed, because sucrose in a batter is also a plasticizer, a humectant, a browning substrate and a regulator of starch gelatinization, and erythritol replaces almost none of that.
The commercial conclusion is uncomfortable but consistent across categories. Moving a brief from 100 percent replacement to 60 or 70 percent frequently removes more sensory problems than every downstream intervention combined, and costs less than all of them. If you have any influence over the target, spend it there before you spend it on masking flavours.
08 Dose
Tolerance, Calculated Per Serving Rather Than Per Day
Erythritol is the best tolerated polyol in commercial use, and the reason is metabolic rather than dose related. Around 90 percent of an ingested dose is absorbed in the small intestine and excreted unchanged in urine, so most of it never reaches the colon to be fermented. Sorbitol, xylitol and maltitol do reach the colon, which is why they cause trouble at much lower doses.
The human data are specific enough to formulate against.
| Study Finding | Figure |
|---|---|
| Single bolus in liquid, healthy adults, no significant symptoms | 20 g and 35 g |
| Single bolus in liquid, healthy adults, increased nausea and borborygmi but not diarrhoea | 50 g |
| Xylitol at the same doses | Significantly worse at every level tested |
| Laxative threshold, adults | 0.66 g/kg bw men, 0.80 g/kg bw women |
| No observed effect level, children aged 4 to 6 | 15 g, equal to 0.73 g/kg bw |
| Children at 20 g and 25 g | Significantly more diarrhoea and severe GI symptoms |
Now the number that constrains real products. EFSA set an acceptable daily intake of 0.5 g/kg body weight per day for erythritol in 2023. For a 70 kilogram adult that is 35 grams per day, across all sources.
A beverage formulated at 5 percent erythritol, which is an ordinary level in a US sugar free soft drink, delivers about 16.5 grams in a 330 ml can. Two cans is 33 grams. One consumer, one day, two drinks, and the ADI is effectively spent before anything else in the diet is counted.
A portfolio level problem, not a product level one
This is why EFSA concluded that both acute and chronic exposure estimates already exceed the ADI across population groups, and why it refused an industry application to exempt erythritol from the laxative warning label. It is also why erythritol dosage should be looked at across a portfolio rather than product by product. A brand selling a sugar free drink, a sugar free ice cream and a protein bar can put a regular consumer over the ADI without any single product looking aggressive.
The steviol glycoside comparison is stark. The ADI for steviol glycosides is 4 mg/kg body weight per day expressed as steviol equivalents, which is 125 times lower as a number but is almost never the binding constraint in practice, because the use levels are so low. One ADI is a real formulation limit. The other rarely is.
09 Safety literature
Where the Cardiovascular Question Actually Stands in 2026
Your customers are asking about this, so a technical document that skips it is not much use. Here is the state of the evidence, presented without a commercial conclusion attached, because the honest answer is that it is unresolved.
The case for concern
A 2023 study in Nature Medicine used untargeted metabolomics in patients undergoing cardiac risk assessment and found circulating erythritol associated with three year risk of major adverse cardiovascular events. Targeted analysis in independent US and European validation cohorts confirmed the association, with adjusted hazard ratios of 1.80 and 2.21 comparing the highest quartile to the lowest. The same group followed up in 2024 with a human intervention showing that a 30 gram dose raised plasma erythritol more than a thousandfold and acutely enhanced platelet reactivity. A 2025 in vitro study reported adverse effects on brain microvascular endothelial cell function at a concentration comparable to a standard sweetened beverage.
The case for caution about that case
Erythritol is produced endogenously in humans through the pentose phosphate pathway, so elevated plasma erythritol may be a marker of dysregulated glucose metabolism rather than a consequence of dietary intake. Two 2023 reviews argued exactly this and noted that most long term animal feeding studies do not support a platelet activation effect. A 2025 review in Cardiovascular Research goes further, summarising work in which erythritol acts as a hydroxyl radical scavenger and shows endothelium protective properties under hyperglycaemic conditions, including a pilot trial in type 2 diabetes where 24 grams acutely improved small vessel endothelial function and 36 grams per day over four weeks reduced central aortic stiffness.
What has not happened
No long term randomised controlled trial of dietary erythritol with cardiovascular endpoints. Until one exists, causality is not established in either direction.
What this means for a purchasing decision
The useful framing is commercial rather than scientific. Whatever the science eventually concludes, the reputational exposure is already real and is already driving reformulation briefs. Some brands have decided to reduce erythritol loading or move it out of front of pack positioning. Others have concluded the evidence does not justify a change. Both are defensible. What is not defensible is a supplier on either side of the trade using an unresolved literature as a selling point, and we are not going to do that with a product we also sell.
10 Regulation
The Same Formula Is Not Legal in Every Market
This section decides more projects than the sensory work does, and it is almost never written down in one place.
| Point of Comparison | European Union | United States |
|---|---|---|
| Erythritol identity | E 968, food additive | GRAS |
| Use in flavoured drinks | Maximum 1.6 percent, that is 16,000 mg/kg, as a flavour enhancer only, and only in energy reduced or no added sugar drinks | Unrestricted as a bulk sweetener |
| Use in other categories | Quantum satis across 66 food categories as a group I additive, for purposes other than sweetening | Unrestricted |
| Erythritol ADI | 0.5 g/kg bw/day, set 2023 | None set |
| Laxative warning | Required above 10 percent added polyols, exemption application refused in 2023 | Not required |
| Erythritol energy value | 0 kcal/g, against 2.4 kcal/g for all other polyols | approximately 0.2 kcal/g |
| Steviol glycoside ADI | 4 mg/kg bw/day as steviol equivalents | Same |
In an EU beverage, erythritol is not legally a sweetener.
It is authorised as a flavour enhancer, capped at 1.6 percent, and restricted to energy reduced or no added sugar drinks. A US formula built on erythritol as the bulk sweetener in a diet soft drink cannot simply be exported. The sweetness has to be reallocated to the glycoside, and the erythritol has to be repositioned as a minor functional ingredient with a different declared purpose on the label.
Outside beverages the EU position is much more permissive, with quantum satis use across most categories, so the constraint is category specific rather than general. Confectionery, bakery and dairy do not face the beverage cap.
The energy value row is worth flagging to marketing teams as well as regulatory. Erythritol is the only polyol assigned 0 kcal/g under the EU nutrition labelling rules. Every other polyol carries 2.4 kcal/g. On a product where the calorie declaration is a claim driver, that difference can be larger than it looks.
11 Supply
What Changed in the Erythritol Supply Chain in 2025 and 2026
Two trade actions have reset the landed cost of Chinese erythritol in the two largest export markets, and any cost model built before 2025 is now wrong.
European Union. Following an investigation opened in November 2023, the Commission imposed definitive anti dumping duties on erythritol originating in China. The duties range from 34.4 percent to 233.3 percent depending on the producer, entered into force in January 2025, apply for five years, and were collected retroactively to June 2024.
United States. Commerce and the International Trade Commission both reached affirmative final determinations, and antidumping and countervailing duty orders on Chinese erythritol took effect on 27 March 2026. The final weighted average dumping margin for the named producers is 85.04 percent, with the China wide entity at 184.26 percent. The petitioner was Cargill.
Verify before you quote
Check the current countervailing rates and your own producer's individual rate in the Federal Register notice before you build a quotation around them, because the published summary tables have contained errors. Duty rates also change on review, and the rate that applies is the one in force on the date of entry, not the date of the purchase order.
Two things follow from this that matter for formulation rather than just for purchasing.
These are trade measures, not safety measures. Nothing in either case concerns product quality or consumer risk.
Confusing the two in a customer conversation will cost you credibility.
Second, the economics of the stevia and erythritol pair have shifted. When erythritol was the cheapest solids in the room, the default was to load it heavily and use the glycoside sparingly. With duties of this magnitude on EU and US bound material, the cost optimum in many formulas has moved toward a higher glycoside contribution and a lower erythritol load, backfilled with fibre, polydextrose or partial sugar retention. That is a formulation change driven by a customs decision, which is not a sentence anyone enjoys writing, but it is where a lot of 2026 reformulation work is actually coming from.
And the obvious escape route is closed in Europe
Allulose is the ingredient most often proposed as an erythritol replacement, because it browns, it behaves like a sugar and it carries no cooling effect. It is not authorised in the European Union. It remains a novel food that has not been added to the Union list, and in June 2025 EFSA concluded that the safety of one applicant's dossier could not be established because the applicant did not supply the additional data requested. That is a data gap rather than a finding of harm, and several further applications are pending, but until one is authorised, allulose cannot lawfully be placed on the EU market. EU formulators who need to reduce erythritol have to work with fibres, polydextrose, isomalt, maltitol or partial sugar retention instead.
12 Applications
Stevia vs Erythritol by Application
Not a category by category introduction, but the failure mode you are likely to hit first and what to do about it.
| Application | What Usually Fails First | Where Erythritol Helps | Where Erythritol Hurts | Sensible First Move |
|---|---|---|---|---|
| Carbonated and still beverages | Thin body, sweetness sitting on top of nothing | Mouthfeel and flavour rounding at low levels, and the cooling effect is largely invisible once dissolved | Legal ceiling in the EU, and ADI exposure at US style levels | Glycoside led sweetness, erythritol at 1 to 2 percent, fibre or gum for body |
| Powdered drinks and tabletop | Volume in the sachet or scoop | This is the core use case, it provides the visible spoonable solids | Cooling is at its strongest in a dry crystalline format | Match particle size to the format, blend with a positive heat of solution carrier |
| Yoghurt and fermented dairy | Body, water holding, syneresis | Solids and water activity control | Cooling perceptible in a cold spoonable product | Combine with fibre or protein, keep erythritol moderate |
| Ice cream and frozen desserts | Hardness and scoopability | Strong freezing point depression from the low molecular weight | Recrystallization during freeze thaw creates grittiness | Balance total freezing point depression deliberately, do not swap gram for gram |
| Cakes, muffins and cookies | Structure, moisture, crust colour | Partial bulk restoration | No browning, no plasticizer effect, drier crumb at high replacement | Cap replacement near 50 percent, add humectant and consider a browning source |
| Chocolate and compressed tablets | Snap, mouthfeel, cooling | Bulk and crystalline structure | Cooling is at its most objectionable here | Consider isomalt or maltitol as the bulk, use erythritol as a minor component |
| Gummies and jellies | Texture, syneresis, crystallization | Solids and water activity | Crystallizes readily, giving grainy texture and cloudiness | Control with doctoring agents such as fibre syrups, watch storage temperature |
| Protein powders and bars | Existing protein bitterness stacking with sweetener bitterness | Bulk and dilution of the glycoside load | Cooling perceptible, cost per unit weight is high at high inclusion | A cleaner glycoside profile plus moderate erythritol usually beats more of either |
13 Specification
What to Settle Before You Request a Quotation
Half of the sensory work R&D teams do in a pilot kitchen could have been avoided at the specification stage. These are the lines that change the outcome.
- On the steviol glycoside side. Which assay the number refers to, since 95 percent Reb A and 95 percent total steviol glycosides are different materials. The full glycoside profile including Reb C and dulcoside A, because minor composition is where the aftertaste differences live. The production route and the applicable E number, which determines labelling and market access. Solubility behaviour at your product's pH and serving temperature, which matters more for Reb D and Reb M than for Reb A. Expected lot to lot variation on the minor glycosides.
- On the erythritol side. Assay and purity, along with residual ribitol and reducing sugars. Particle size distribution and crystal habit, which is the specification line most buyers omit and which directly controls dissolution rate, perceived cooling intensity and powder flow. Coarse granular material gives the spoonable sugar look a tabletop product needs. Finely milled material dissolves faster and reduces the cooling sensation in dry mixes and tablets. Moisture. Country of origin and HS classification, because that is what determines your duty exposure in the EU and the US. And, for EU destined beverages, confirmation of the declared functional purpose your customer intends to use.
- On a blend. If you are buying a pre blended system rather than the two components, ask for the ratio, whether the blend is dry mixed or co processed, the sucrose equivalence at a stated use level rather than a headline multiplier, and enough material to test in the finished application. A one gram sample supports a water tasting, and a water tasting is not a purchasing decision.
The grades and assay bases we hold are listed on the stevia extract page.
14 Decision framework
The Decision, Stated Plainly
If your product needs sweetness and nothing else, the glycoside can carry most of it and you may need very little erythritol.
If your product needs volume, solids or a visible portion, erythritol is doing work no high intensity sweetener can do, and the question is how much rather than whether.
If your product needs structure, browning or moisture retention, neither ingredient solves it and you should be looking at replacement level, humectants and fibre before you look at sweeteners at all.
If the sensory defect is bitterness or lingering sweetness, that is a glycoside problem and adding erythritol is an indirect fix at best.
If you are shipping to Europe, check the category rules before the formula, not after.
And if your cost model for erythritol predates 2025, rebuild it.
We supply steviol glycoside grades across the Reb A, Reb D and Reb M range, erythritol, and blended systems of the two built to a stated sucrose equivalence rather than a headline ratio. Tell us the application, the sugar replacement level you are targeting and the destination market, and we will propose a starting ratio and ship enough material to test it in your own matrix rather than in water. Samples are free, MOQ is 1 kg, and quotations are returned within 24 hours.
15 Questions
Frequently Asked Questions
Why is erythritol listed first on the ingredient panel of most stevia and monk fruit sweeteners?
Because ingredients are declared in descending order by weight, and erythritol is usually more than 95 percent of the blend. The steviol glycoside is often under 2 percent. A consumer reading the label concludes the product is mostly erythritol with a little stevia added, which is factually correct and frequently surprises the brand owner. If that ordering is a marketing problem for your positioning, the options are to reduce the erythritol fraction and accept a smaller serving format, or to reposition the product away from a cup for cup claim.
Is stevia or erythritol cheaper per unit of sweetness?
Steviol glycosides are dramatically cheaper per unit of sweetness delivered and dramatically more expensive per kilogram. Comparing kilogram prices is meaningless. Calculate cost per unit of sucrose equivalence at your actual use level, then add the cost of whatever you have to add to replace the bulk. Once duties are included on EU or US bound erythritol, the ranking in a full formula can invert relative to what it was two years ago.
Can a stevia and erythritol blend be used cup for cup like sugar?
Yes, and that is exactly what most retail blends are engineered to do. It requires the erythritol fraction to be dosed for volume rather than sweetness, with the glycoside making up the sweetness shortfall, and it requires attention to bulk density so that a cup of the blend weighs what the recipe expects. Note that cup for cup solves the volume problem only. It does not restore browning, moisture retention or structure, so baked results still differ from the sugar version.
Does erythritol survive baking, pasteurisation and retort?
Yes. Erythritol is thermally stable across normal food processing conditions and does not participate in Maillard browning, which is precisely why baked goods made with it stay pale. It is also stable across a wide pH range, so acid beverages are not a problem. Steviol glycosides are likewise heat stable, though their behaviour in prolonged low pH storage should be verified for your specific shelf life.
Can stevia and erythritol be labelled as natural?
It depends entirely on the market and on the production route, and the answer is not the same for the two ingredients. Erythritol is produced by fermentation of glucose, which supports a natural positioning in many markets but not automatically in all of them. For steviol glycosides, the route determines the additive identity in the EU, and a fermentation derived Reb M carries a different E number than leaf extracted material. Get written confirmation of the route from your supplier before any claim is drafted, because this is a common source of late stage label rework.
Does erythritol raise blood glucose or insulin?
No. Erythritol is absorbed and excreted without being metabolised for energy, and human studies consistently show no meaningful effect on blood glucose or insulin. Steviol glycosides likewise do not raise blood glucose. This is one of the few areas where the two ingredients behave the same way, which is why blends of them are common in diabetic and ketogenic product positioning.
What can we use if a customer wants erythritol removed entirely?
There is no single replacement, which is the point of this article. Depending on which function you need back: soluble fibres such as chicory inulin, resistant dextrin or polydextrose for bulk and body, isomalt or maltitol for confectionery structure, glycerin for humectancy, and partial sugar retention where the brief allows it. Allulose is the closest functional match because it browns and does not cool, but it is not authorised in the EU. Expect to use two or three ingredients where erythritol was one, and expect the cost to rise.
Is monk fruit plus erythritol different from stevia plus erythritol?
Functionally the erythritol side is identical. The difference is on the sweetener side, in sensory profile, cost per unit of sweetness and regulatory identity. Do not assume that combining monk fruit with stevia gives you a better result than either alone, though, because receptor level work found no synergy between rebaudioside M and mogroside V and suggested they compete for the same binding sites. Blending two natural high intensity sweeteners is a decision to be tested rather than assumed.
16 Evidence base
References
[1] Belloir, C., et al. (2024). A receptor-based assay to study the sweet and bitter tastes of sweeteners and binary sweet blends: the SWEET project. Chemical Senses. DOI: 10.1093/chemse/bjae041
[2] Heikel, B., Krebs, E., Köhn, E., Busch-Stockfisch, M. (2012). Optimizing synergism of binary mixtures of selected alternative sweeteners. Journal of Sensory Studies, 27(5), 295–303. DOI: 10.1111/j.1745-459X.2012.00396.x
[3] Andersen, B. V., et al. (2023). Temporal check-all-that-apply on the sensory profiling of sucrose-replaced sweetener blends of natural and synthetic origin. Journal of Sensory Studies. DOI: 10.1111/joss.12838
[4] Gao, J., Brennan, M. A., Mason, S. L., Brennan, C. S. (2017). Effects of sugar substitution with Stevianna on the sensory characteristics of muffins. Journal of Food Quality. DOI: 10.1155/2017/8636043
[5] Storey, D., Lee, A., Bornet, F., Brouns, F. (2007). Gastrointestinal tolerance of erythritol and xylitol ingested in a liquid. European Journal of Clinical Nutrition. PMID 16988647
[6] EFSA Panel on Food Additives and Flavourings (2023). Re-evaluation of erythritol (E 968) as a food additive. EFSA Journal, 21(12). DOI: 10.2903/j.efsa.2023.8430
[7] Witkowski, M., et al. (2023). The artificial sweetener erythritol and cardiovascular event risk. Nature Medicine, 29, 710–718. DOI: 10.1038/s41591-023-02223-9
[8] Commission Regulation (EU) 2015/1832 of 12 October 2015, on the use of erythritol (E 968) as a flavour enhancer in energy-reduced or with no added sugars flavoured drinks. View source
Reviewed by
Prof. Bingji MaScientific Advisor at PhytoVale. Professor and medicinal plant science expert. Research focus covers Chinese medicinal materials, phytochemistry, natural products and medicinal plant quality.
Written by
FloraContent Editor at PhytoVale. Product content and editorial specialist, focused on botanical ingredient specifications, sourcing information and B2B technical content.
This article is intended for ingredient sourcing, formulation and technical information only. It does not constitute medical, regulatory or legal advice. Permitted uses, maximum levels, labelling requirements and duty rates change, and should be verified against current legislation for the destination market and against the current published rates for your own supplier and HS classification at the time of purchase.
Need the Right Stevia Grade?
Tell us your application and sweetness target. We’ll recommend a suitable glycoside profile and provide sample availability, COA and pricing.
