How Gem Index estimates a price

A plain-English walk-through of the model — what each factor does, why coloured stones resist a single price grid, and how to read the range you get back.

The core model

Gem Index uses a multiplicative model rather than an additive one, because gemstone value factors compound rather than stack. The engine starts from a base price per carat for the species and multiplies it by a series of factor weights:

estimated price/ct = base/ct × colour × clarity × transparency × cut × treatment × origin × certification × size

Each factor is a multiplier centred near 1.0. A factor above 1 lifts value; below 1 discounts it. The output is expressed as a range, not a point, because real trade prices for identical-looking stones vary with negotiation, market timing and buyer type.

Base price per carat

Every supported variety carries a seed base reflecting typical wholesale value for a clean, well-cut, average-colour stone of that species. This is the anchor the multipliers act on. Bases differ enormously — a fine ruby's base is many times a citrine's — which is why species selection is the single most important input.

Colour

For most coloured stones colour is the dominant driver. The engine separates tone (how light or dark) and saturation (how vivid), because the ideal is species-specific: medium-to-medium-dark with high saturation wins for sapphire and ruby, while some pastels are prized in other species. Over-dark or greyish/brownish colour is penalised.

Clarity & transparency

The model applies clarity expectations by GIA type. Type I stones (e.g. aquamarine, tanzanite) are expected eye-clean, so visible inclusions cost more. Type II (ruby, sapphire, spinel) tolerate minor inclusions. Type III (emerald, red tourmaline) are expected to be included, so clarity is weighted more gently. Transparency — from crystalline to opaque — is a separate multiplier.

Cut & finish

Cut quality, symmetry and polish affect how much of the colour and brilliance the stone returns face-up. A poorly cut stone with windowing or extinction is discounted even if the rough was fine.

Treatment

Treatment logic is adapted per species. Heated versus unheated corundum, oiling grades in emerald, and the untreated-by-default status of spinel and tsavorite are each modelled differently. Where a treatment is standard and accepted, the discount for "treated" is mild; where untreated material is genuinely rare (fine unheated sapphire, natural-colour emerald), the premium for "none" is large. Aggressive treatments such as glass-filling collapse value.

Origin

For species where origin carries a market premium — Kashmir and Burma sapphire, Burmese ruby, Colombian emerald, Brazilian Paraíba — the engine applies an origin multiplier. Origin only matters when it is credibly documented, so its weight is tied to certification.

Certification

A major-lab report reduces buyer risk on species, treatment and origin, and the market pays for that certainty. The model lifts value modestly for major-lab certification, less for domestic labs, and treats uncertified fine stones more conservatively.

Size

Price per carat rises with size for fine material, because large clean crystals are rarer. The size curve is non-linear and species-specific — it climbs far more steeply for ruby, alexandrite and Paraíba than for amethyst.

How to read the result

Treat the midpoint as a considered central estimate and the range as the realistic spread between a motivated seller and a careful buyer at wholesale. The figures are indicative and in USD. Before any real transaction, confirm species, treatment and origin with an accredited laboratory — see our disclaimer.