Belvedere's Median Price Is Arguing With Itself

How many sales does it take before a number can be trusted? Ten? Fifty? In most Marin cities, that threshold clears itself every month without anyone noticing. In Belvedere, it doesn't clear at all, and the market's own public data proves it.

As of the August 2026 update, the same market-tracking page that publishes Belvedere's numbers reported the average sale price at $5.28 million for the trailing month, down 16.6 percent from a year earlier. Scroll down the same page and the trailing three-month median, covering May through July 2026, shows home prices up 46.5 percent year over year, at a median of $7.5 million. Same city. Same source. Same month of publication. Two directions.

That isn't a data error. It's what happens when a real estate market is small enough to fit in a school gymnasium.

The Same Data, Two Different Stories

Belvedere sold 12 homes in July 2026, up from 8 the year before. A move from 8 to 12 sales sounds like growth. In a market the size of San Rafael or Mill Valley, it would be a rounding error. In Belvedere, it's close to half the month's entire transaction count shifting on its own.

Average and median measure different things, and that gap is exactly where the trouble lives. Average gets pulled hard by one very large or very small sale. Median resists that pull, up to a point, but with only a handful of closings in any given window, even the median is standing on thin ice. When the sample is 8 to 12 homes, one $15 million estate or one $2 million fixer can swing either number by double digits without the underlying market having moved at all.

Zoom out to a full calendar year and the picture steadies. Across 2025, Belvedere closed 29 homes at an average price of roughly $6.3 million, or about $1,900 a square foot. That twelve-month figure is a far more honest read of where the market actually sits than any single month or quarter. The problem is that most buyers and sellers don't wait twelve months to check a portal. They check it the week they're deciding whether to make an offer, and that week's number might be telling them something that isn't true.

Thirty Sales a Year Is Not a Sample, It's a Neighborhood Roster

Belvedere is half a square mile of land holding around 2,100 residents, and it typically closes somewhere between 25 and 35 home sales a year. That works out to two or three closings a month, some months fewer. A single listing sitting on the market for a long stretch can distort the days-on-market figure by itself. Earlier this year, Belvedere's median days on market spiked to 244 in January, a number driven almost entirely by one trophy-tier listing that hadn't found a buyer yet. The same market-tracking page's August 2026 update showed July's trailing average time on market at just 27 days, down from 69 days a year earlier, a swing that has far more to do with which specific homes were on the market than with any citywide shift in buyer appetite.

This is the part a national valuation estimate can't see. It treats every closed sale as one data point among a large, stable pool, the same math that works fine in a market with a few thousand annual transactions. Belvedere doesn't have a few thousand. It has dozens. When the pool is that small, the identity of the seller, the specific street, and even who happens to be shopping that quarter can matter more than any citywide trend line.

What One Buyer Did to Golden Gate Avenue

The clearest illustration of this happened in late August 2025, when former Apple design chief Jony Ive bought four homes on Belvedere Island in a single package for a combined $73 million. Three of the four sit next to each other on Golden Gate Avenue, all built in the early 1900s and ranging from roughly 3,600 to 8,900 square feet. The fourth, on Beach Road, was built in 1999 and comes with hillside elevators and private dock access.

The largest of the four, a property known as Locksley Hall, sold for $43.5 million, a slight discount from the roughly $47.5 million it fetched about a decade earlier. The second Golden Gate Avenue home, the Organ House, closed at $14.62 million, and the third at $6.9 million. The Beach Road property sold for $7.98 million. Across the package, Ive paid between $1,858 and $4,888 per square foot depending on the house, a range wide enough that no single number from that transaction could stand in for the neighborhood.

Local luxury broker Barbara Major, commenting on the sale to the San Francisco Chronicle, put it plainly: homes in Belvedere generally

"don't sell for those prices."

She pointed to four other Belvedere sales that had closed since March 2025, all in the more typical $7 million to $9 million range. That's the honest ceiling for most of the market. The Ive package sat well above it, and if those four sales land inside the same twelve-month window that a portal uses to calculate an annual average, they alone can pull the whole city's number upward by a meaningful margin.

The Fifty Million Dollar Listing Still Sitting in Inventory

Belvedere doesn't need another whale transaction to prove the point again. As of April 2026, the priciest active listing in Marin County sat at 20 Crest Road, an 8-bedroom, 11-bath compound on a terraced 40,000-square-foot lot, asking $50 million. The home was designed by architect William Wurster, whose other work includes the 52-story tower at 555 California Street in San Francisco and Case Study House No. 3 in Los Angeles, with landscape architecture by Thomas Church. It last changed hands in 1979 for $1.2 million.

Whenever that listing closes, whatever price it closes at, it will do to Belvedere's next reporting period exactly what the Ive package did in 2025. One sale, one number, and a citywide statistic that moves because of a single closing rather than because buyers and sellers as a group changed their behavior.

How to Actually Read a Belvedere Comp

A single median or average can't carry the weight buyers and sellers want it to carry here. What holds up better is reading the market by the sub-area a specific property sits in, since each behaves with its own logic:

• Belvedere Island's estate row, including Golden Gate Avenue and the hillside streets above it, where recent sales have run from the high single digits into the tens of millions, and where privacy and off-market handling matter as much as square footage.
• Belvedere Lagoon, where waterfront access and dock condition drive price more than lot size, and where the state-regulated bulkhead calendar adds its own timeline to any transaction.
• Beach Road, home to the small number of condominiums that offer the most accessible entry point into the city, though they trade so rarely that a buyer waiting for one to list needs real patience.
• Corinthian Island and the Cove, generally the more moderate end of Belvedere pricing, still commanding a premium for the address but without the trophy-estate ceiling found on the Island proper.

Treating these as one market and averaging across all of them is exactly how a $73 million package deal or a $50 million listing ends up doing the talking for a city that mostly transacts in the $4 million to $8 million range.

What This Means at the Table

If you're comparing Belvedere to another Marin city using a single portal number, you're comparing noise to signal. The fix isn't complicated, but it does require going past the headline figure. Ask for the actual list of closed sales in the specific sub-area you're evaluating, not the citywide roll-up. Look at a twelve-month window rather than a single month or quarter, since Belvedere's tiny sample size means short windows swing hard in either direction. And if a portal's estimate seems out of step with what similar homes nearby have actually closed for, trust the closed comps over the algorithm. In a market this small, the algorithm doesn't have enough sales to learn from.

FAQ

Is "Belvedere" the same market as "Belvedere Tiburon" on real estate portals? No. Some data providers combine Belvedere and neighboring Tiburon under a single "Belvedere Tiburon" label, which reports a different median, a different inventory count, and a different sales pace than Belvedere on its own. Confirm which label a number is using before comparing it to anything.

How many homes actually sell in Belvedere in a typical year? Roughly 25 to 35, based on recent annual totals, with 2025 closing at 29 sales. That's the full context needed before trusting any single month's figure.

What's the single best question to ask an agent about a Belvedere comp? Which specific streets and which specific sales they pulled the comp from, and over what window. In a market this size, the answer to that question matters more than the number itself.

Belvedere rewards buyers and sellers who read past the headline. If you're weighing a purchase or a sale here and want the actual comps behind the number, not the portal's best guess, First California Realty can walk through them with you. Let's connect, starting with a confidential home valuation.