Published 2 November 2026 · By Jatin Dua, Broker of Record, RE/MAX Quantum Realty

Our AI condo estimator starts from a price per square foot for your area, learned from 8,406 Toronto and GTA condo sales, then adjusts for size, floor, view, exposure, age, condition, parking and your building’s own sales record. Tested on 1,673 sales it had never seen, its median error was 7.5%.
This post is for Toronto and GTA condo owners who want to know what sits behind the number my free AI condo estimator gives them. I built the tool to be quick to use and honest about its limits. Below I walk through the data, each step of the formula and how we tested it, in plain words.
What data is the AI condo estimator built on?
The estimator learns from sold prices, not asking prices. On 22 September 2026 we recalibrated it from REALM, the research tool that draws on TRREB MLS sold data. The set holds 8,406 condo-apartment sales that had a recorded size range. They sold between 22 March and 22 September 2026, across Toronto and eight GTA cities: Mississauga, Brampton, Vaughan, Markham, Richmond Hill, Oakville, Burlington and Milton.
Six months of sales are not all equal. A sale in March happened at a different price level from a sale in September, and condo prices were still easing by about 0.8% a month over that window. So every sale was time-adjusted to September 2026 before the model learned from it. The code notes that the whole process should be re-run every quarter, so the model keeps pace with the market.
Two more data sets sit beside the sales:
- a building index of 2,732 GTA buildings that had REALM sales between September 2025 and September 2026;
- a rent model built on 16,218 condo apartments leased between 22 June and 22 September 2026, which gives the “rent instead” figure in your report.
For market context, the sales the model learned from came from a soft market. TRREB MLS data via REALM shows condo-apartment medians for January to October 2026 below the same period of 2025 in all eight GTA cities the tool covers outside Toronto, and in most Toronto districts. Mississauga was down 9.3%, Markham 13.2% and Oakville 6.0%. A model that learned from older prices would overshoot in a market like this. That is why the time adjustment matters.
How does the estimator turn sales into a price per square foot?
Each of the 18 areas in the tool has its own base price per square foot. The areas run from Toronto Central/Downtown, North York and Scarborough to Humber Bay Shores, Mimico, The Queensway and the eight cities above. Each base is set at a reference size of 700 square feet.
Your unit’s size is then applied through a size exponent of 0.33. In plain words, price per square foot falls as units get bigger. A 1,400 square foot suite does not sell for twice what a 700 square foot suite sells for in the same area. The exponent bends the line so it matches how different sizes actually traded.
Every setting in the model, from the area base prices to the floor premium, was fitted on the tool’s own formula using a method called robust least squares. Ordinary least squares lets a few odd sales, such as a penthouse or a rushed sale, pull the whole line toward them. Robust methods reduce that pull, so one strange sale cannot reset the price for a whole area.
How do floor, view, exposure, age and condition change a condo valuation?
Once the model has a base price for your size and area, it adjusts for what makes your unit different. These are the settings it learned from the sales.
| Feature | How the model treats it |
|---|---|
| Floor | About +0.37% per storey above a typical comparable floor (the 11th). Premium capped at +14%, discount floored at −6%. |
| View | From +2% (open city) to +8% (full lake), measured against the local norm. |
| Exposure | South earns a small premium, east a small discount; north and west are neutral. |
| Building age | Multipliers fitted from sales: brand new carries the highest premium, 30+ years the lowest. |
| Condition | Renovated about +4.5%; dated about −7%. |
| Parking | A dollar amount: about $47,000 at a reference price level, scaled to the area. |
| Locker | A dollar amount: about $13,500. |
Source: the AI condo estimator’s calibration on 8,406 REALM (TRREB MLS) condo-apartment sales, 22 March to 22 September 2026.
Take a unit on the 25th floor. It is 14 storeys above the typical comparable floor, so the floor setting on its own adds roughly 5%. A unit on the 4th floor sits seven storeys below, so it takes a small discount, and never more than 6%.
There is one more safety step. When several adjustments stack up, say a high floor, a full lake view, a south exposure and a fresh renovation, the model compresses the total so it cannot run away. That keeps a pile of premiums from adding up to a price no buyer in that area has paid.
Why are condo parking and lockers added in dollars, not percentages?
Parking and lockers are added as dollar amounts because, as the tool’s code puts it, that is how condos actually trade. A parking spot is not worth more because the suite above it is bigger. A buyer is paying for the spot itself.
So the model adds about $47,000 for owned parking at a reference price level, then scales it with the area’s price level. A downtown Toronto spot is worth more than one in a lower-priced area. An owned locker adds about $13,500.
If you want the market side of parking, views and floors, my guide to the three things that move a Toronto condo price covers it. This post sticks to how the model prices them.
Try it on your own unit now. The estimator below takes about 45 seconds over three short steps, and your report shows each of these settings at work.
Condo Valuation
What’s your condo
worth today?
Three quick steps. Condos don’t price like houses — your floor, your view, and whether you own parking move the number more than anything else. This weighs all of them.
Reading recent condo sales…
Estimated market value
—
$0–$0
Most likely $0 · about $0 per square foot
What moved the number
Starting from what comparable units in your area sell for, here’s what your specifics added or subtracted.
Market context
—
—
Two units, same floor plan,
$90,000 apart.
That happens constantly in condos — one has the parking, the right exposure, or a board that keeps the reserve fund healthy. A model can’t see your status certificate. I can.
How does the estimator avoid double counting what is normal in your area?
A lake view in Humber Bay Shores is not the same thing as a lake view in North York. In Humber Bay, a lake view and a parking spot are the norm, not an upgrade, so the comparable sales there already include them. Adding a full premium on top would count them twice.
To stop that, the model knows what is already normal in each area’s sales: the share of comparable sales that came with parking or a locker, and the usual view. It then prices only the difference between your unit and that local norm.
This works in both directions. Where most sold units had parking, a unit without a spot is priced below the local base. Where few sold units had parking, owning one adds closer to the full amount. The same logic applies to views. A standard view in an area where most sales had open views pulls the number down a little, while a full lake view in an area where that is rare pushes it up.
How does building-level matching make the number sharper?
Two condos with the same size, floor and view can sell for quite different prices because they are in different buildings. Finishes, amenities, management and reputation all show up in what a building’s units sell for. An area average misses this.
So the estimator has a building layer. The building index built from REALM sales between September 2025 and September 2026 covers 2,732 GTA buildings. It gives each building a premium or a discount against what the model would otherwise predict. When you pick your building from the address box, the tool also fills in your area, the building’s age and its usual parking.
Two guard rails keep the building layer honest:
- A building with only a few sales has its premium shrunk toward zero, because two or three sales can be a fluke.
- Every building’s adjustment is capped between −20% and +25%.
When your building does have sales on record, your estimated range is tighter. If your building is new or has had no recent sales, the model leans on the area level and your report shows a lower confidence level. My post on pricing a condo when nothing in your building has sold explains how I handle that case by hand.
How do maintenance fees factor into a condo valuation?
Maintenance fees are judged against the typical fee for a building of the same age. Older buildings genuinely carry higher fees, and many of them are all-inclusive. Marking down every older building for a higher fee would be unfair to their owners.
Only a fee that is high for its vintage nudges the value down, and even then the nudge is light. The sold data showed that once the model knows the building, the fee barely moves the price. You can also pick “not sure” for the fee and still get your estimate. For the buyer’s side, including how fees affect what a buyer can borrow, read do high maintenance fees lower my condo’s value.
How was the model tested on sales it had never seen?
A model can look perfect on the sales it learned from and still fail on new ones. The fair test is a holdout. You set aside sales the model never sees while it is being fitted, then compare its estimates with what those units really sold for.
We held back 1,673 sales for this test. Here is what it showed:
- Calibration cut the area model’s median error from 18% to 9%.
- Adding the building layer then took the median error from 8.2% to 7.5%, with 97% of the test sales matched to a building.
- Overall, 61% of the estimates landed within 10% of the actual sale price.
A 7.5% median error means half the estimates were within 7.5% of the real price and half were further off. On a $600,000 condo, 7.5% is $45,000. That is useful for planning. It is not tight enough to set a list price on its own, and I would rather tell you that up front.
I built this estimator to give you a fast, honest starting number and to show its working. The holdout results are strong for a free tool, and I am proud of the building layer in particular, because it is what lets the model tell two similar buildings apart. Still, about 4 in 10 estimates land more than 10% from the sale price. Use the number to plan, then let me check it against the things a model cannot see before you make a big decision.
What can a condo valuation model not see?
Even a well-tested formula only knows what you enter and what past sales show. It cannot read your building’s reserve fund, its maintenance-fee history, any special assessment or the status certificate. It cannot see the quality of your renovation, a tired kitchen or a layout that wastes space. It also struggles with one-of-a-kind suites, such as a large penthouse. Those are the cases where automated models miss most, as I explain in why online home estimates break down at the top end.
That is why every report says plainly that it is an automated estimate for information only. It is not an appraisal and not a Comparative Market Analysis (CMA). If you want to know how those three differ and when each one counts, my post comparing an AI valuation, a CMA and an appraisal lays it out.
The model was calibrated to September 2026 price levels and is meant to be re-run each quarter. If condo prices keep easing, a September-level figure can run a little high by the time you sell. Do not use the estimate as a value for a lender, a court or a tax filing. Ask for a written CMA, and get a formal appraisal where a lender or lawyer needs one.
What to do next
- Run your unit through the free AI condo estimator, and pick your building from the list so the building layer can work.
- Read the “What moved the number” section and check that each input matches your unit.
- Note your confidence level. If it says “Indicative”, treat the range as a rough guide.
- Gather your latest status certificate or fee statement and a list of any renovations.
- Ask me for a written CMA before you list, refinance or settle an estate.
Need help with this? Talk to me.
I’m Jatin Dua, Broker of Record at RE/MAX Quantum Realty, and I’m the best agent to do this kind of work for you in Toronto and across the GTA. Over $100 million in GTA sales and 100+ families helped sell. Feel free to contact me any time: call or text 833-330-1925 or email [email protected].
Frequently asked questions
How does an online condo estimator work?
An online condo estimator starts from what comparable condos in your area have sold for, then adjusts for your unit’s own features. The AI condo estimator on jatindua.com uses a price per square foot for each of 18 GTA areas, learned from 8,406 sales in 2026, then adjusts for size, floor, view, exposure, age, condition, parking, locker and your building’s own sales record. In 2027 it gives a range, a most likely figure and the reasons behind it.
How are condo values calculated in Toronto?
Toronto condo values are calculated mainly from recent sales of similar units, adjusted for differences in size, floor, view, parking and condition. My AI condo estimator does this with a formula fitted by robust least squares on Toronto and GTA sales from March to September 2026. A realtor’s written CMA then adds what a formula cannot see, such as the status certificate, the finish of the unit and the current competing listings.
What data does the jatindua.com condo estimator use?
The estimator uses sold data from TRREB MLS, accessed through REALM. Its 22 September 2026 calibration used 8,406 condo-apartment sales from 22 March to 22 September 2026 across Toronto and eight GTA cities, all time-adjusted to September 2026. It also uses a building index of 2,732 GTA buildings and a rent model built on 16,218 leased condo apartments from mid-2026.
Does the AI condo estimator use current listings?
No, the estimator is calibrated on sold prices, not on current listings or asking prices. Its 2026 calibration used sales from 22 March to 22 September 2026, adjusted to September price levels, and it is designed to be re-run every quarter. For a price that reflects the listings you would compete against in 2027, ask for a written CMA.
How much does a parking spot add to a condo’s value in the estimator?
The estimator adds owned parking as a dollar amount of about $47,000 at a reference price level, scaled up or down with your area’s price level. A downtown Toronto spot therefore counts for more than one in a lower-priced GTA area. Where parking is already normal in local sales, as in Humber Bay Shores, the model only prices the difference from that norm.
How much is a higher floor worth in a Toronto condo?
In my 2027 AI condo estimator, each storey above the 11th floor adds about 0.37% to the estimate, so a 25th-floor unit gains roughly 5% from height alone. The floor premium is capped at +14%, and the discount for low floors never goes past −6%. These settings were fitted on 2026 sales across Toronto and the GTA.
Why does a lake view add less to a condo in Humber Bay Shores?
A lake view adds less in Humber Bay Shores because it is the norm there, so comparable sales already include it. The AI condo estimator knows the usual view and the share of sales with parking in each GTA area, and it prices only the difference between your unit and that local norm. This avoids counting the same feature twice.
What is a holdout test for a home value estimator?
A holdout test checks a value model on sales it never saw while it was being built. For my AI condo estimator, 1,673 GTA condo sales were held back. On those sales the median error was 7.5%, and 61% of estimates landed within 10% of the actual sale price. It is the fairest way to judge a model before you rely on it in 2027.
Do maintenance fees lower the estimated value of my condo?
Only slightly, and only when your fee is high for a building of its age. The AI condo estimator compares your fee with the typical fee for that vintage, because older GTA buildings often carry higher, all-inclusive fees. The 2026 sold data showed fees barely move price once the building is known, so the nudge is light.
Is an AI condo estimate the same as an appraisal?
No. An AI condo estimate is an automated figure for information only. It is not an appraisal and not a Comparative Market Analysis (CMA), and it cannot read your building’s reserve fund, fee history, special assessments or status certificate. In Toronto and the GTA in 2027, use it to plan, then ask for a written CMA, or an appraisal if your lender or lawyer needs one.
Sources
- Jatin Dua: What Is My Condo Worth? Toronto and GTA Condo Value Calculator
- Wikipedia: Training, validation, and test data sets
- Wikipedia: Robust regression
- ICE Mortgage Technology: What is an Automated Valuation Model (AVM)?
- Jatin Dua: What Is My Condo Worth? Floor, View and Parking
- Jatin Dua: How to Price My Condo When Nothing in My Building Has Sold
- Jatin Dua: Do High Maintenance Fees Lower My Condo’s Value?
Related reading
- AI Condo Estimator Features Explained: Floor, View, Parking, Confidence and the ‘What Moved the Number’ Breakdown (2027)
- 8 Times to Use an AI Condo Estimator in 2027 (and When You Need a CMA Instead)
- How Accurate Is an Online Condo Value Estimate? What Our 1,673-Sale Test Shows (2027)
- Try the Free AI Condo Estimator: What You Get, What Happens Next and Your Privacy (2027)
About the author — Jatin Dua, Toronto and GTA real estate broker
I am Jatin Dua, Broker of Record and co-founder of RE/MAX Quantum Realty Inc., Brokerage, Unit 101, 799 The Queensway, Toronto. I work with buyers and sellers across Toronto and the GTA and have helped more than 100 families sell. Four-plus years of active GTA transactions and over $100 million in sales volume.
Reach me at [email protected] or 833-330-1925, or book a call.
More from the library
- Buying a Luxury House in Toronto: The Complete 2027 Guide
- Best Realtor in Richmond Hill (2027): How to Choose One, With Real Numbers
- Lincoln Lakefront Homes: Beamsville, Vineland and Jordan Harbour
- Downsizing From a Toronto Luxury Home: The Math Most People Get Wrong
- How Much Is My House Worth in the East End (Leslieville, the Beaches, East York)? August 2026 Prices
- O2 Maisonettes, 220 George St: Two-Storey Homes in a 50-Suite Building
How this estimate is produced
The estimate blends recent selling prices for comparable units in the area with the details you entered: size, bedrooms, floor, exposure, parking, locker and condition. It is a statistical estimate, not an appraisal and not a formal opinion of value.
What it cannot see
- Interior condition and upgrades beyond what you told it
- Unit-specific issues: noise, an obstructed view, a pending special assessment
- Your reserve fund and where the maintenance fee is heading
- What is competing with you in your own building right now
What moves the number most
- Floor and exposure, which can separate otherwise identical units
- Parking and locker, each priced separately by buyers
- Kitchen and bathroom condition
- Maintenance fee per square foot against the rest of the building
Treat it as a range, not a price
A single figure to the dollar would be false precision. Use the estimate to decide whether a move is worth exploring. Before you price anything for the market, the number needs checking against what is actually selling and sitting in your building this month, and that takes a person. I will do that for you at no cost and with no obligation.
Market data is sourced from the Toronto Regional Real Estate Board. Information is deemed reliable but is not guaranteed. This tool does not provide an appraisal or a formal opinion of value, and it is not a substitute for advice on your specific property. Jatin Dua, Broker of Record, RE/MAX Quantum Realty.
