Vake costs $2,945 a metre and Vashlijvari $1,069 — in the same city

In April 2026, 3,641 new-build apartments sold in Tbilisi, up 10.7% on the same month a year earlier, at an average of $1,408 per square metre. The two ends of the price range were Vake at $2,945 and Vashlijvari at $1,069, with Mtatsminda ($2,881) and Chugureti ($2,235) at the top and Didi Dighomi ($1,136), Gldani ($1,135) and Samgori ($1,124) at the bottom. But volume did not follow price in either direction: Saburtalo led with 760 sales, then Didi Dighomi with 732 and Samgori with 571, while Vake managed 128 and Mtatsminda 66.

Half the transactions — 51.8% — were below $1,100 per square metre, and apartments of 51 to 80 square metres accounted for 53.5% of sales. Above $3,000 per square metre the premium segment was 1.5% of the market. The picture that falls out of this is of a city whose sales engine is mid-sized flats in outer districts, with the prestige addresses contributing headlines and very little turnover. An owner deciding whether to sell should know which of those two markets their flat is actually in.

Georgian citizens took 75% of purchases. Of the foreign quarter, Israeli buyers were 11% of all sales, Russian buyers 3%, and other nationalities 11% between them. That composition matters for anyone trying to read risk: a market carried by domestic income and domestic lending responds to Georgian interest rates and wages, not primarily to sentiment in one or two source countries. It also means a shock to any single foreign buyer group removes a slice of demand rather than the floor under it.

An average per square metre is a blend of finished and shell units, of ground floors and top floors, of a tower beside a main road and a courtyard block behind it. Two flats on the same street can sit either side of a district average by 30% on condition alone. Use the table to see which direction a district is priced in and how deep its sales are, then price your own flat from comparable finished units with the same view, floor and access — and treat a single month's figure as one reading, not a trend.