Steviol Glycoside Purity Grades Explained for Food Manufacturers
Steviol glycoside purity grades such as 80%, 90%, 95%, 97%, 98% and 99% describe how much of a defined glycoside or group of glycosides is present in an ingredient. They are not a universal scale from “industrial grade” to “premium food grade.” A product marked 95% total steviol glycosides (TSG) is not the same specification as 95% rebaudioside A (Reb A) or 95% rebaudioside M (Reb M). The percentage only becomes meaningful once the ingredient identity and test basis are known.
For the same individual glycoside, a 98% grade contains three more percentage points of the target compound than a 95% grade at those nominal assay values. A 99% grade contains one point more than a 98% grade. Whether either difference improves a drink, dairy product, bakery filling or tabletop sweetener depends on the other compounds present, the physical form of the powder and the finished formulation. Purity alone does not predict taste, solubility or legal suitability.
This guide compares the grades food manufacturers actually see in quotations, examines published sensory and processing research, and shows how to select a workable grade without paying for a percentage that has no demonstrated benefit in the finished food. It then explains the essential regulatory and COA checks. For a broader comparison of glycoside identities and application selection, see our stevia grade selection guide.
01 · Grade comparison
Steviol Glycoside Purity Grades: 80%, 90%, 95%, 97%, 98% and 99%
The first mistake in a purity comparison is to treat all percentages as measurements of the same substance. Commercial stevia quotations may describe mixed leaf extracts, highly purified Reb A, Reb D or Reb M, or materials obtained through different manufacturing processes. A numeric grade can refer to one molecule or the sum of several. The table below describes what each percentage can mean; it does not assign a universal food-approval status or a guaranteed application to any grade.
| Quoted purity | What the number means when the assay is defined | What it does not establish | Practical purchasing interpretation |
|---|---|---|---|
| 80% or 90% | 80% or 90% of the stated target glycoside(s), on the supplier's specified assay basis | That the material qualifies for use as a food additive in the destination market | Request the exact glycoside profile, intended use, regulatory identity and whether the item is a concentrated ingredient or an intermediate |
| 95% | At least 95% of a named single glycoside, or at least 95% of a defined group if labelled TSG | That 95% TSG contains 95% Reb A or Reb M | Common reference point for many high-purity regulatory specifications; confirm precisely which substances count toward the 95% |
| 97% | A tighter commercial assay limit for the same defined substance or mixture | A separate internationally recognised quality class | Compare its price and lot assay with 95% and 98% rather than assuming it belongs to a different regulatory category |
| 98% | More of the same target compound than 95%, if identity and test basis match | Better taste, faster dissolution or a higher legal maximum use level | Evaluate the composition of the remaining fraction and test performance in the actual product |
| 99% | One percentage point more target compound than 98%, on a like-for-like basis | A globally defined “pharmaceutical” or “premium food” grade | Useful only where tighter composition control or another documented process benefit justifies the premium |
To compare grades fairly, keep three things constant: the target molecule or permitted glycoside sum, the measurement basis and the manufacturing or formulation conditions. Reb A 98% is not an upgraded version of TSG 95%. They may have fundamentally different dominant sweet compounds. Conversely, Reb A 95% and Reb A 98% are reasonable candidates for a like-for-like comparison if both assays refer to Reb A as a mass fraction on the same basis.
Fast purchasing answer
If the quotation states only “stevia extract 98%,” it is incomplete. Ask whether 98% refers to total steviol glycosides, Reb A, Reb D, Reb M, or another defined mixture. That answer is more important than the two-digit purity number.
Where do 80% and 90% grades fit beside 95% to 99%?
Lower-assay grades are not automatically poor ingredients, and higher-assay grades are not automatically better food additives. A Reb A 80% powder may have a substantial remainder consisting of stevioside and other steviol glycosides. A Reb A 98% powder leaves much less room for those compounds. That can create a real difference in flavour or formulation behaviour, but the outcome depends on the specific remainder, not merely the size of the remainder. A mixed TSG 90% extract poses another question: which glycosides make up that 90%, and does its particular identity satisfy the food-additive specification in the intended market?
| What the buyer has been offered | What can reasonably be concluded | What must be resolved before choosing |
|---|---|---|
| Reb A 80% versus Reb A 98% | The same target molecule is present in different proportions, assuming equivalent assay methods | Actual glycoside profiles, lawful intended use, equal-active-dose flavour testing and the cost of each formulation |
| TSG 90% versus TSG 95% | One declared sum is higher; the main glycoside may still differ | The identity and composition of the counted glycosides, and whether either material meets the relevant food-additive definition |
| Reb M 95% versus Reb M 98% or 99% | The target Reb M fraction is progressively tighter on a comparable basis | Whether less secondary material produces a demonstrated sensory, handling or consistency benefit worth the premium |
This is why a manufacturer should not begin by assigning each grade to an application such as “95% for bakery” or “99% for clear beverages.” First define the molecule and lawful ingredient identity; then determine whether the narrower composition of the higher grade provides a measurable advantage. For a lower-assay extract, a separate safety or regulatory assessment may be needed rather than simply increasing the dose to compensate for its content.
A commercial grade is not the exact result for every production lot
A supplier may sell a “Reb M 95% grade” under a written specification of Reb M ≥95%. One batch might test at 95.4%, another at 97.6%, and both can meet that same commercial specification. These numbers are illustrative, not PhytoVale batch data. In the same way, a material sold as “Reb M 98% grade” normally communicates a higher minimum specification, not a guarantee that every lot measures exactly 98.00%. Food manufacturers should distinguish the product grade, the contracted minimum and the actual assay when comparing quotations.
02 · What is being measured
95% TSG, 95% Reb A and 95% Reb M Are Three Different Specifications
Total steviol glycosides (TSG) describes a sum. An individual-glycoside assay identifies one component. Consider a hypothetical leaf-extract powder consisting, on the same dry basis, of 79% Reb A, 11% stevioside and 5% other glycosides included in the declared assay. Its total is 95%, but its Reb A content is 79%. It would be misleading to offer that powder as Reb A 95% simply because the sum reaches 95%.
A separate Reb M product with 95% Reb M could contain small amounts of other glycosides. Those secondary compounds will depend on how the material was made and purified. A higher Reb M number usually leaves less room for other components; it does not identify those components, nor does it prove that they would have caused poor performance. This distinction matters especially when moving between leaf-extracted blends, purified individual glycosides and enzymatic or fermentation-derived preparations.
| Example supplier description | Measured target | What the buyer should compare |
|---|---|---|
| Steviol glycosides 95% | Defined sum of eligible glycosides | Individual profile and identity of the authorised mixture |
| Reb A 95% or 98% | Rebaudioside A itself | Reb A mass fraction, secondary-glycoside profile and test basis |
| Reb D 95% | Rebaudioside D itself | Reb D assay, other components and formulation performance |
| Reb M 95%, 98% or 99% | Rebaudioside M itself | Reb M assay, origin/route, solid form and dissolution behaviour |
| Glucosylated steviol glycosides | Defined parent and glucosylated glycosides | Special compositional rules and the relevant dried, dextrin-free calculation where applicable |
Dry-basis purity and the powder arriving at the factory
A dry-basis value removes moisture from the assay denominator. If a powder is 95% of a specified glycoside on a dry basis and contains 3% moisture, its approximate glycoside fraction in the as-received material is 95% × 97% = 92.15%, assuming the two figures use consistent mass definitions. A 100 kg delivery therefore does not contain 95 kg of the target on an as-received basis in that example. This is a mass-accounting calculation, not evidence that the batch fails a specification expressly defined on a dry basis.
Glucosylated steviol glycosides require particular attention because the relevant specification may calculate purity on a dried, dextrin-free basis. That number should not be treated as the amount of glycosides per kilogram of a commercial powder containing dextrin. The actual delivered composition and water content must be considered separately. JECFA and EU specifications provide the technical basis for these distinctions.[1][2]
Purity of an ingredient is not the concentration of a finished sweetener blend
A common confusion arises when a processor buys a high-purity glycoside but receives, or makes, a formulated tabletop powder containing carriers. If 10 kg of a Reb A ingredient assayed at 95% on an as-received basis is blended with 90 kg of other declared ingredients, the 100 kg blend contains approximately 9.5 kg Reb A, or 9.5% Reb A by mass. The source ingredient can legitimately be described as “Reb A 95%”; the completed blend cannot be described as 95% Reb A. Its composition, carrier identity and labelling must be stated separately.
For food manufacturers, this distinction affects the addition rate, ingredient declaration and economic comparison. Two suppliers can offer the same 95% source-grade specification yet deliver very different commercial preparations once carriers or other sweeteners are added. Obtain the formulation-level composition before translating a raw-material assay into grams per serving.
03 · Taste evidence
Does 98% or 99% Purity Taste Better Than 95%?
Not necessarily, and the published evidence does not support a general rule that each purity increase produces a cleaner taste. Sensory studies clearly show differences between glycoside molecules. They are much less conclusive about a strict 95%-versus-98% comparison of the same glycoside when all other variables are controlled. Confusing these two questions leads to oversimplified grade recommendations.
In a 2020 consumer study, Tao and Cho asked 126 participants to evaluate 0.1% solutions of Reb A, Reb D and Reb M against sucrose. At the concentrations tested, Reb A showed greater bitterness, while Reb D and Reb M had sensory characteristics closer to sucrose. Reb D and Reb M also had more lingering sweetness than sucrose. The study compared different glycosides; it did not demonstrate that Reb M 98% tastes better than Reb M 95% in a finished beverage.[3]
| Sweetener in the 2020 sensory test | In-mouth bitterness (15-cm scale) | Bitterness one minute after tasting |
|---|---|---|
| 14% sucrose reference | 0.8 | 0.4 |
| 0.1% Reb A | 3.5 | 1.6 |
| 0.1% Reb D | 1.1 | 0.6 |
| 0.1% Reb M | 1.0 | 0.6 |
These are the published mean scores, not measurements made by PhytoVale. The study used fixed solution concentrations, not matched-sweetness finished products, and the sucrose reference was at 14%. It therefore illustrates meaningful glycoside identity differences but cannot serve as an estimate of how a 95% grade compares with a 98% grade of the same molecule.[3]
Processing research shows why the rest of the composition still deserves attention. A 2014 analysis of four commercially processed stevia extracts found 23.7 mg of minor natural glycosides per 100 mg extract in the membrane-purified material, compared with an average of 2.4 mg per 100 mg in ethanol-crystallised extracts. A trained panel of 13 people found the highest liquorice intensity in the membrane-purified material. Yet two ethanol-crystallised samples with differing chemical compositions did not show significant sensory differences. Composition can matter, but every measured difference is not automatically perceptible.[4]
For food formulation, the important sensory question is not “Which percentage is highest?” It is “At equal target sweetness, does the more expensive grade perform differently in this recipe?”
Directly relevant evidence: Reb A 40% versus 95% in peach juice
A separate published study tested stevia extracts standardised to 40% and 95% Reb A in peach juice. Both were used at 0.1055 g/100 mL of beverage. In the time–intensity analysis, the peak bitterness score was 6.43 for the 40% material versus 5.87 for the 95% material, and the integrated bitterness response was 143.33 versus 126.02. Those differences supported a lower bitterness response for the higher-Reb-A extract under the study conditions.[12]
| Peach juice experiment | 40% Reb A extract | 95% Reb A extract | Meaning for manufacturers |
|---|---|---|---|
| Peak bitterness (time–intensity score) | 6.43 | 5.87 | Higher-Reb-A extract reduced the observed bitterness peak |
| Integrated bitterness response | 143.33 | 126.02 | The 95% extract produced less total measured bitterness across time |
| Consumer overall liking (9-cm line scale) | 3.24 | 3.27 | No meaningful improvement in overall acceptance was observed |
This distinction is important. Although bitterness was reduced, overall liking stayed low for both stevia samples. In addition, the researchers added the two commercial extracts at the same powder weight, not at equal Reb A mass, so the study does not isolate purity from target-glycoside dose and secondary composition. It supports a difference between these particular 40% and 95% preparations in peach juice. It does not prove that moving from Reb A 95% to 98% or 99% produces a perceptible or commercially useful improvement. Any such claim still requires a direct, controlled application trial.[12]
What a controlled 95%–98% pilot comparison should look like
Start with the same target glycoside, for example Reb A 95% and Reb A 98%, and use amounts adjusted to provide the same measured mass of Reb A. A direct equal-weight comparison would introduce a predictable active-content difference before the sensory trial even begins. Then prepare both samples in the intended food matrix, with the same pH, process, flavour system and storage conditions. Compare sweetness onset, bitter or liquorice notes, sweetness linger and overall acceptance using an appropriately designed blind test.
The results can justify moving to a higher grade, staying with a lower grade or specifying a different glycoside blend altogether. The molecule can matter more than the extra three purity points: a formulation struggling with the bitterness of Reb A 98% may need a different glycoside strategy rather than Reb A 99%. Readers comparing Reb A with Reb M can use our separate Reb A vs Reb M guide.
04 · Manufacturing performance
How Purity Grades Affect Dissolution, Stability and Food Processing
A high assay is useful when a manufacturer needs predictable loading of a particular glycoside, but it does not tell the formulator whether that powder will dissolve at the required concentration or remain stable through processing and storage. Solid-state structure, crystal form, temperature, pH, other glycosides and the food matrix also influence behaviour. A grade may pass its chemical assay and still be a poor choice for a particular production line.
Reb M: a real example of purity and solubility being different questions
Prakash and colleagues reported that purified crystalline Reb M (>95%) dissolved to approximately 0.1 g per 100 mL in water at 25°C after five minutes, whereas their amorphous material exhibited a reported solubility of 1.1%–1.3% at 25°C. The study also reported an equilibrium solubility of about 0.26% at 25°C. These figures represent different material states or test conditions and should not be collapsed into one interchangeable specification.[5]
The lesson for purchasing is straightforward: “Reb M 98%” does not specify the crystal form, particle properties or dissolution protocol. A 98% grade cannot be assumed to dissolve faster than a 95% grade without a like-for-like test. This is particularly relevant when an operator prepares a concentrated pre-solution, processes a cold beverage, or expects a clear finished drink. In such cases, record dissolution time, clarity, sediment after cooling and any changes during storage rather than approving a sample solely because the COA reads 98%.
Purity at receipt is not stability after processing
Research on Reb A in acidified beverages demonstrates that degradation during storage depends on conditions and formulation ingredients. Gelinas and co-workers identified chemical degradation routes in acidic drinks and observed that caramel colour and certain flavour components altered the formation of different degradation products. This does not imply that every stevia beverage is unstable; it shows that ingredient interactions and storage conditions can matter independently of the starting assay.[6]
A 99% grade at receipt is not a substitute for a shelf-life test. For acidic beverages, manufacturers should evaluate the actual pH, storage temperature, packaging and ingredient matrix. For hot-filled products, clarity at filling is not sufficient if particles can form after the liquid cools. For dry blends, the concern may instead be dose uniformity and segregation. These are different formulation problems and cannot be solved by purity alone.
Application-by-application grade selection
| Food format | First decision | How purity matters | What still needs a pilot test |
|---|---|---|---|
| Clear RTD beverages, flavoured water and tea | Choose an acceptable glycoside or blend for the desired flavour | Tighter individual assay can help control composition; it does not guarantee clear dissolution | Target sweetness, cold-water dispersion, haze, acid stability and shelf life |
| Coffee and other hot beverages | Evaluate sweetness in the beverage's real bitterness and aroma matrix | 95% versus 98% changes active mass at equal powder weight | Hot dissolution, sweetness linger and taste at actual serving temperature |
| Yoghurt, dairy drinks and plant-based alternatives | Control sweetness after acidification or fermentation | Specified glycoside and minor-component profile may matter more than nominal assay alone | Sensory balance, interactions with proteins, pH changes and storage |
| Bakery, fillings and confectionery | Account for sweetness as well as the functional bulk removed with sugar | Higher purity reduces ingredient mass needed for a defined glycoside dose but does not replace sugar's bulk | Thermal process, colour, texture, moisture and final sweetness |
| Tabletop powders and dry premixes | Determine target sweetness per sachet or serving | A very pure glycoside can be difficult to dose uniformly at tiny quantities without a carrier | Blend uniformity, flow, segregation, humidity sensitivity and label composition |
None of these categories requires 99% purity simply because it is a beverage, dairy food or bakery product. The practical approach is to establish a sensible grade and then prove that it works under the conditions of manufacture. Product teams that need more detailed applications can consult our food and beverage grade selection guide; the present article remains focused on what the purity number adds to that decision.
For a like-for-like purity trial, normalise both powders to the same mass of the specified glycoside before comparing flavour, clarity or storage. If the manufacturing line first makes a concentrated premix, test that premix separately from the final diluted beverage; a material that performs acceptably at serving concentration may still create processing difficulties in a much more concentrated intermediate. Finally, distinguish a chemical assay improvement from a manufacturing improvement: lower dosing weight is calculable, but lower haze, shorter dissolution time or reduced rework must be measured.
05 · Cost in use
When Is 98% or 99% Steviol Glycoside Purity Worth the Premium?
Comparing quoted price per kilogram without correcting for the target glycoside can lead to a poor purchasing decision. But correcting for purity is only the first step, not a substitute for sensory or production trials. A useful screening calculation is quoted price per kg ÷ assay fraction, provided the products are the same molecule and the assays are reported on the same basis.
Consider three fictional offers for the same individual glycoside, with consistent assay definitions. These are teaching figures, not current PhytoVale prices.
| Grade | Illustrative quoted price | Price per theoretical kg of target glycoside | Premium versus 95% by target content |
|---|---|---|---|
| 95% | $80.00/kg | $80 ÷ 0.95 = $84.21 | Reference |
| 98% | $84.00/kg | $84 ÷ 0.98 = $85.71 | More expensive in active-content terms |
| 99% | $85.00/kg | $85 ÷ 0.99 = $85.86 | More expensive in active-content terms |
Using the same assumptions, a 98% grade can command up to 98/95 − 1 = 3.16% more per kilogram than a 95% grade before it becomes more expensive per unit of the target compound. The equivalent threshold for 99% versus 98% is only 99/98 − 1 = 1.02%. Those thresholds are arithmetic, not established market price differentials. If the quotation premiums exceed them, the buyer needs some additional, demonstrated reason to choose the higher purity.
How much powder is actually needed for the same target dose?
Suppose the specification calls for 100 mg of a particular glycoside in a pilot formulation. On the same assay basis, 95% powder would require approximately 105.26 mg, 98% powder 102.04 mg, and 99% powder 101.01 mg. The difference between using 95% and 98% powder is about 3.22 mg of ingredient for each 100 mg of target glycoside. Whether that saving matters depends on production scale, quote difference and formulation performance.
Recalculate using the actual lot assay if that is how the purchase contract and formulation are managed. Moisture can change the as-received calculation, and the price comparison should use the same basis for both suppliers. If the higher grade gives materially better clarity, fewer flavour corrections, lower rework or more consistent production, these benefits may justify a premium; none is guaranteed by the purity number itself.
What the percentage difference means for a 1,000-litre production run
For a plant-scale example, assume a trial formulation requires 100 mg/L of one specified glycoside, or 100 g of that compound in 1,000 L of finished beverage. Assume the quoted assays refer to the as-received mass of comparable powders and no additional process loss. At that same glycoside target, the differences in material consumption become concrete:
| Purity of the same glycoside | Powder per 1,000 L | Powder cost using the illustrative quotes above | Difference versus 95% grade |
|---|---|---|---|
| 95% | 105.26 g | $8.42 | Reference |
| 98% | 102.04 g | $8.57 | 3.22 g less powder, but $0.15 more per 1,000 L |
| 99% | 101.01 g | $8.59 | 4.25 g less powder, but $0.17 more per 1,000 L |
The 98% to 99% step saves only about 1.03 g of powder per 1,000 L under these assumptions. This does not make 99% unnecessary; it shows how small the mass saving can be before considering a possible taste or processing benefit. The 100 mg/L target is strictly a calculation example, not a recommended formulation level or confirmation of permitted use in any food category. Actual loading and legal limits must be checked separately, including steviol-equivalent rules when relevant.
06 · Food additive identity
Do Food Regulations Require 95%, 98% or 99% Purity?
Many high-purity steviol glycoside specifications use a 95% threshold, but there is no single worldwide rule saying that every material below 95% is illegal or that 98% and 99% constitute separate regulatory approvals. Legal suitability depends on the defined additive, manufacturing process, specified assay and permitted use in the intended market. Commercial labels should not be substituted for official specifications.
| Example regulatory identity | Relevant purity approach | What the 95% does not mean |
|---|---|---|
| EU E 960a, steviol glycosides from stevia | Specified steviol glycosides from leaf extraction, with a defined total assay | That one glycoside such as Reb A makes up 95% of the product |
| EU E 960b(i), Reb M from specified fermentation | At least 95% for the defined combined Reb M/D/A/B fraction under that particular specification | That every permitted batch contains at least 95% Reb M alone |
| EU E 960c, enzymatically produced steviol glycosides | Composition and production details depend on the precise sub-entry; some specify an individual Reb M assay | That a Reb M product made by any enzyme route is automatically interchangeable under one sub-entry |
| EU E 960d, glucosylated steviol glycosides | Specified parent and glucosylated glycosides on the applicable dried, dextrin-free basis | That 95% of every as-shipped powder kilogram consists of glycosides |
These are abbreviated identity distinctions, not a substitute for the full legislation. The consolidated EU specification regulation and later amendments control the exact compositional, microbiological, solvent and production conditions.[1][7][8] A particularly relevant recent development is Commission Regulation (EU) 2026/2309, published on 19 October 2026, addressing a specified enzymatic Reb M route. As of this article's 23 October 2026 update, its application date has not yet arrived. Its existence illustrates why confirming the production route can matter as much as confirming the assay.[8]
In the United States, the FDA has evaluated many GRAS notices for high-purity steviol glycosides with a minimum purity of 95%, and has not questioned the submitters' GRAS conclusions for the specified substances under their stated conditions of use. That is not a general FDA approval for anything labelled “stevia 95%.” The agency also distinguishes highly purified steviol glycosides from stevia leaves and crude stevia extracts, which are not considered GRAS for use as sweeteners.[9]
A practical distinction for international manufacturers
Purity approval is not use-level approval. Where the relevant rules express sweetener limits as steviol equivalents, the permitted amount must be calculated from the specific glycoside composition and applicable conversion factors, not from a generic “95% powder” figure. Detailed EU and US use rules are covered in our steviol glycoside regulation guide.
Why two 95% products can give different steviol-equivalent calculations
Consider 100 mg (dry basis) of a powder with 95% Reb A and another with 95% Reb M. The target glycoside in each case is 95 mg, but the steviol-equivalent conversion factor differs because the two molecules have different molecular weights. The commonly specified rounded factors are approximately 0.33 for Reb A and 0.25 for Reb M.[1][13]
| Component-only illustration | 95% Reb A powder | 95% Reb M powder |
|---|---|---|
| Powder amount (dry basis) | 100 mg | 100 mg |
| Named glycoside present | 95 mg Reb A | 95 mg Reb M |
| Rounded conversion factor | 0.33 | 0.25 |
| Steviol equivalents from the named 95% fraction | 31.35 mg | 23.75 mg |
The table calculates only the identified 95% fraction; it does not calculate the total steviol equivalents of either powder. Other steviol glycosides in the remaining 5% must also be included using their respective factors when applicable. Nor are these amounts measures of perceived sweetness. Food manufacturers need the complete component profile to calculate a compliant formula, not merely the headline 95% grade.
07 · Purchase decision
How Food Manufacturers Should Choose and Verify a Purity Grade
The most defensible decision starts with the sweetener molecule and the application, not with the maximum available assay. Select the glycoside or blend that delivers the required sweetness, check that the production route and additive identity are suitable for the destination market, and then compare two or more grades of that same material in the actual food. Only after this is the price premium for 98% or 99% meaningful.
For example, a manufacturer making a clear flavoured water may begin with Reb M samples of two commercial grades, adjust the sample weights to the same target Reb M mass, and test clarity and taste under intended storage conditions. A manufacturer of a dry tabletop blend may instead find that dose uniformity and carrier design dominate the engineering decision. Neither project benefits from assuming that 99% is automatically the safest or best-performing option.
The minimum COA information needed to compare grades
The supplier's COA should give the named glycoside or defined TSG sum, the actual result for the lot, the reporting basis and an appropriate analytical method. A more detailed glycoside profile is useful where secondary components may affect the desired formula. A well-designed HPLC method can separate and quantify multiple compounds, but an unqualified chromatogram peak-area percentage must not be assumed equivalent to a calibrated mass assay. Liu and colleagues validated a rapid HPLC-UV method for 13 steviol glycosides and reported an apparent Reb M value of 100.8% in one raw-ingredient measurement. That is an analytical result, not a physically possible material mass fraction above 100%.[10]
Evidence from actual products reinforces the need to verify a claim. Gardana and Simonetti analysed 40 commercial stevia extracts, seven sweeteners and three stevia-sweetened beverages; nine extracts and one beverage did not satisfy the EU regulation applicable to that work. The sample set is not an industry-wide failure-rate estimate, but it shows why supplier specifications and independent testing may matter for commercial procurement.[11]
| If your situation is... | Choose this next action | Why |
|---|---|---|
| The quotation says “98% stevia” without naming the assay | Clarify TSG versus single-glycoside identity before comparing prices | The headline percentages may describe different materials |
| Both offers are Reb A 95% and Reb A 98% | Correct dose for active content and run a blinded application test | Higher assay alone does not prove a noticeable sensory benefit |
| A clear drink uses crystalline Reb M | Check solid form, dissolution, haze and storage rather than only COA purity | Practical solution behaviour depends on more than chemical assay |
| A 99% grade costs significantly more | Calculate the 1.02% purity-adjusted price threshold versus 98%, then test any claimed benefit | A one-point assay gain may not justify a larger premium |
| The formula will be marketed in the EU or US | Confirm the precise regulatory entry or relevant GRAS basis and intended use | 95% does not itself establish legal eligibility |
A final decision table: stay at 95%, move to 98%, or specify 99%?
Once the regulatory identity and glycoside type have been settled, a higher grade should be justified against a defined outcome. The table provides a decision framework rather than pretending that one purity level suits every food.
| Observed procurement and formulation result | Practical decision |
|---|---|
| 95% and 98% of the same glycoside meet all applicable specifications, with equivalent sensory and production results | Keep the qualifying grade with the lower validated cost in use; a higher number alone is not a reason to switch |
| 98% gives a reproducible flavour or clarity improvement in a blinded, equal-active-dose test | Consider 98% if the measurable benefit outweighs the ingredient and validation premium |
| 99% produces no improvement over 98%, and only slightly reduces powder use | Do not pay an additional premium solely for the one-point assay gain |
| Lower-assay 80% or 90% material seems more economical | Verify the complete glycoside profile and regulatory identity before any pricing or dosage comparison |
| TSG 95% and Reb M 95% are both offered for the same formulation | Choose the appropriate molecule or mixture first; identical percentage labels do not make them substitutes |
| Highest-assay powder causes concentrated-premix haze or poor reconstitution | Reassess physical form and processing conditions instead of automatically demanding even higher chemical purity |
The result should be a purchasing specification that names the target glycoside, the minimum assay and reporting basis, the applicable quality parameters, the intended manufacturing route where relevant, and the application tests the grade must pass. Then a lower-priced 95% grade can be selected when it meets the same finished-product requirements, while a higher grade can be specified where its benefit has been measured rather than assumed. For available commercial specifications and sample discussion, see the PhytoVale stevia extract page.
08 · Frequently asked questions
Frequently Asked Questions About Steviol Glycoside Purity
What does 95% steviol glycoside purity mean?
It means at least 95% of the compound or defined combination of compounds specified by the product's assay, on the stated reporting basis. A 95% TSG extract measures a defined total; Reb A 95% or Reb M 95% refers to the individual glycoside. Confirm which definition the supplier uses.
Is 98% stevia extract better than 95% for food manufacturing?
It has a greater proportion of the target compound if both products are otherwise comparable and measured on the same basis. That can affect dose and composition, but does not guarantee better taste, solubility or storage stability. Compare equal target doses in the finished food.
Is 99% Reb A or Reb M worth paying more for?
Only if tighter composition control or an observed processing or sensory benefit justifies the higher cost. Compared with 98%, the extra one percentage point gives an approximately 1.02% price-premium threshold on a purely active-mass basis, before considering any other differences.
Can food manufacturers use 80% or 90% steviol glycosides?
There is no universal answer based on purity alone. Such products may be commercial mixtures, intermediates or preparations subject to different legal definitions. Their use in food requires evidence that the specific composition, process and intended application satisfy the destination market's rules; many high-purity leaf-derived sweetener specifications use a 95% minimum.
What is the difference between TSG 95% and Reb M 95%?
TSG 95% measures an eligible sum of steviol glycosides. Reb M 95% measures Reb M as one molecule. A high TSG result does not establish that Reb M represents the same proportion of the powder, and the two products cannot be treated as compositionally equivalent.
Does higher purity make Reb M more soluble?
Not as a general rule. Published research shows that crystalline and amorphous Reb M can exhibit different dissolution behaviour. Comparing commercial 95% and 98% grades requires controlled testing of physical form, concentrations, processing conditions and the specific secondary-glycoside composition.
What should a steviol glycoside COA report?
At minimum, the batch identity, target analyte or defined total, measured assay, reporting basis and analytical method should be clear. A detailed glycoside profile, moisture results and relevant quality tests can help explain differences between grades. These records support ingredient assessment; they are not a substitute for confirming regulatory status and final-product performance.
References
Primary Studies and Official Specifications
Identity and regulatory specifications
[1] European Commission. Commission Regulation (EU) No 231/2012, specification entries for steviol glycosides (E 960a–E 960d), as amended. Official EUR-Lex text.
[2] JECFA. Steviol glycosides: specifications and analytical methods, including the 96th JECFA meeting amendments (2023). WHO JECFA database.
Sensory evidence and processing behaviour
[3] Tao R, Cho S. Consumer-Based Sensory Characterization of Steviol Glycosides (Rebaudioside A, D, and M). Foods. 2020;9(8):1026. doi:10.3390/foods9081026. Full-text original study.
[4] Espinoza MI and colleagues. Identification, Quantification, and Sensory Characterization of Steviol Glycosides from Differently Processed Stevia rebaudiana Commercial Extracts. J Agric Food Chem. 2014;62:11797–11804. doi:10.1021/jf502878k. PubMed record.
[5] Prakash I, Markosyan A, Bunders C. Development of Next Generation Stevia Sweetener: Rebaudioside M. Foods. 2014;3:162–175. doi:10.3390/foods3010162. Full-text original study.
[6] Gelinas BS, Tello E, Peterson DG. Mechanisms of Rebaudioside A Degradation and Ingredient-Sweetener Interactions in Beverages during Storage. Molecules. 2022;27:1385. doi:10.3390/molecules27041385. Full-text original study.
EU and US regulatory distinctions
[7] European Commission. Commission Regulation (EU) 2025/652 of 2 April 2025. Amended specifications addressing Reb M produced by fermentation using Yarrowia lipolytica. Official regulation.
[8] European Commission. Commission Regulation (EU) 2026/2309 of 16 October 2026. Specific enzymatically produced steviol glycoside specifications; published 19 October 2026. Official regulation.
[9] US Food and Drug Administration. Aspartame and Other Sweeteners in Food. Section on certain steviol glycosides and GRAS notices. FDA information.
Analytical methods and commercial market samples
[10] Liu Z and colleagues. Rapid and Economic Determination of 13 Steviol Glycosides in Market-Available Food, Dietary Supplements, and Ingredients: Single-Laboratory Validation of an HPLC Method. J Agric Food Chem. 2020;68:10142–10148. doi:10.1021/acs.jafc.0c03453. PubMed record.
[11] Gardana C, Simonetti P. Determination of steviol glycosides in commercial extracts of Stevia rebaudiana and sweeteners by ultra-high performance liquid chromatography Orbitrap mass spectrometry. J Chromatogr A. 2018;1578:8–14. doi:10.1016/j.chroma.2018.09.057. Original study.
Direct application research and steviol-equivalent factors
[12] Medeiros A, Tavares E, Bolini HMA. Descriptive Sensory Profile and Consumer Study: Impact of Different Nutritive and Non-Nutritive Sweeteners on the Descriptive, Temporal Profile, and Consumer Acceptance in a Peach Juice Matrix. Foods. 2022;11(2):244. doi:10.3390/foods11020244. Compares peach juice sweetened with extracts containing 40% versus 95% Reb A; reports quantitative bitterness and consumer liking, with different glycoside doses at equal extract addition. Full-text original study.
[13] European Commission. Commission Regulation (EU) 2022/1922. Steviol glycoside specifications and conversion factors, including Reb A 0.33 and Reb M 0.25. Official EUR-Lex text.
Written by
Xinjuan XuAssociate Professor, Ph.D. · Scientific Advisor, PhytoVale. Research focus: stevia agronomy, crop physiology, leaf quality and steviol glycoside accumulation.
This article is intended for business-to-business ingredient evaluation and food-formulation reference. Numerical quotation, purity and dosage examples are illustrative, not PhytoVale batch results or current prices. Commercial grade, food-additive identity and finished-food suitability must be evaluated for the specific formulation, manufacturing route and destination market. Ingredient manufacturers and finished-product companies remain responsible for applicable regulatory compliance, product tests and labelling. © 2026 PhytoVale.
Compare the Right Purity Grades for Your Formula
Tell us which glycoside you need, the destination market, your intended application and the grade range you want to evaluate. PhytoVale can discuss available Reb A, Reb D and Reb M specifications and appropriate sample documentation before your pilot trial.
State the target glycoside, requested assay and application to help us match the grade.
