There was almost nothing for sale in the Castro and Eureka Valley in June. One condo and two single-family homes were on the market at the end of the month. Both are the lowest June counts in our ten-year data.
What happened to prices from there is different for each property type, which is worth walking through separately.
Data as of July 27, 2026, from the San Francisco MLS. This report covers June 2026. Some recent-month figures may still be updated as late sales are reported.
The MLS groups this part of the city into a district called Eureka Valley/Dolores Heights, and that is the data behind this report. It is a larger area than the Castro alone. Along with the Castro it takes in Dolores Heights, Corona Heights, much of Duboce Triangle, and the blocks running up toward Upper Market and Twin Peaks.
So the numbers here describe a fairly varied set of homes: flats and condos near the commercial streets, and larger single-family houses up in the hills. That range matters when you read a median, and we will come back to it.
A note on method first. This district closes about five condo sales in a typical month, so we lean on the three-month figures throughout. June was busier than usual with 12 sales.
Inventory is the headline. There was 1 condo actively for sale at the end of June. The next lowest June in our ten-year data was 2, in 2021. A more ordinary June here has somewhere between 6 and 14.
Months supply came in at 1.0 on a three-month basis, down from 2.6 a year ago. Months supply estimates how long it would take to sell everything currently on the market at the current pace of sales, and lower means a tighter, more competitive market.
Competition followed. Over the past three months, 83% of condos sold above their asking price, up from 57% a year ago. The typical sale closed at about 114% of its original list price, compared with 98% last year, so sellers went from accepting slightly less than their first asking price to clearing it by a wide margin. Homes went under contract in a median of 21 days, down from 31.
Prices, though, are down. The three-month median is $1,325,000, about 12% below the same period last year.
Those two facts sit oddly together, and we would not want to over-explain them. At eight sales a quarter, which particular condos close has a lot of say in where the median lands. A run of smaller one-bedrooms will pull it down even in a month when every one of them drew multiple offers. The competition measures and the median are describing different things here, and right now they disagree.
What we would take from it: buyer demand in this district is clearly strong, and the price you should expect depends much more on the specific unit than on the neighborhood figure.
| Metric | June 2026 | 3-month | vs. year ago* |
|---|---|---|---|
| New listings | 3 | 7.7 avg | −15% |
| Active listings | 1 | 5.3 avg | −56% |
| Pending sales | 11 | 8.7 avg | +24% |
| Closed sales | 12 | 8.0 avg | +14% |
| Median sale price (12 closed sales) | $1,375,000 | $1,325,000 | −12% |
| % sold over list | 83.3% | 83.3% | +26.2 pts |
| % of original price | 110.8% | 113.5% | +15.2 pts |
| Days on market | 24 | 21 | −10 days |
| Months supply | 0.2 | 1.0 | −62% |
Same caution here, more so. This district closes about five single-family homes a month, and June had 5. We lean on the three-month figures, and even those rest on 22 sales.
The three-month median sale price is $4,181,500. That is the highest reading in our ten-year data, and it is up about 42% from $2,943,000 for the same three months last year.
A 42% jump in a year is a very large move, and at this sales volume we would hold it loosely. A handful of large houses in the Dolores Heights hills closing in the same quarter can lift the median well beyond what any individual home did. The direction is real. The precise size of it is less certain than the number looks.
Supply is as thin as we have seen it. Two single-family homes were actively for sale at the end of June, the lowest June in our ten-year data. Months supply is 0.3 on a three-month basis, which matches the lowest reading in our ten-year data, down from 2.0 a year ago.
One number moved against the grain, and it is worth flagging rather than skipping. The share of single-family homes selling above asking actually fell, from 93% a year ago to 77%. At the same time, the typical sale finished at about 123% of its original list price, up from 111%.
Read together, that suggests a market where the homes that do attract competition are being bid up very hard, while a slightly larger share are selling at or near their asking price. That is what you would expect when sellers start pricing more ambitiously. It is not a sign of weakening demand.
| Metric | June 2026 | 3-month | vs. year ago* |
|---|---|---|---|
| New listings | 5 | 6.7 avg | +25% |
| Active listings | 2 | 1.3 avg | −85% |
| Pending sales | 3 | 6.0 avg | +13% |
| Closed sales | 5 | 7.3 avg | +47% |
| Median sale price (5 closed sales) | $5,400,000 | $4,181,500 | +42% |
| % sold over list | 80.0% | 77.3% | −16.0 pts |
| % of original price | 127.4% | 122.6% | +12.0 pts |
| Days on market | 5 | 11 | −3 days |
| Months supply | 0.4 | 0.3 | −85% |
A couple of the terms above, briefly. % sold over list is the share of homes that sold above their asking price. % of original price compares the final sale price to the very first asking price, so it also captures any price cuts along the way. Days on market is how long a typical listing took to go from hitting the market to accepting an offer. Pending means the seller accepted an offer, and closed means the sale finished, which usually happens about a month later because of escrow.
And the median: it is the middle value. Half the sales were above it, half below. We use it instead of the average because one very expensive or very cheap sale can't throw it off. In a district with this much variety in housing, that helps, though as noted above it does not solve everything.
Market data is useful. We genuinely enjoy this stuff. But it is one part of a much bigger picture, and it only goes so far.
This district is a good illustration. A one-bedroom flat off Castro Street and a four-bedroom house on a Dolores Heights cul-de-sac are both in these numbers, and neither one is described well by a figure that averages them together. The single-family median moved 42% this year. Almost no individual home did that.
Perfectly timing the market is something that really only happens in hindsight. What matters far more is your own situation: where you are in your life, what you are trying to do next, and what your options actually look like right now.
If you own here and are wondering what the past year has done to your value, the answer depends on which of these two markets your home sits in, and on the particulars of your block and your building. We are glad to work that out with you.
That is the work we do with buyers and sellers every day, and it is why we would rather talk with you about your specifics than hand you a chart and wish you luck.
If you are curious what all this means for you, get in touch and we'll talk it through. You can also read more about how we work with buyers, how we work with sellers, or Eureka Valley and Dolores Heights and the other San Francisco neighborhoods we cover. If you'd like to hear how this goes from people who have been through it with us, our client testimonials are worth a few minutes.
And if you have already done your research and think you would like to work together, the fastest way to start is to fill one of these out:
We would love to hear from you. You can also meet the team if you'd like to know who you'd be working with.
Castro and Eureka Valley market report by Danielle Lazier - Vivre Real Estate, top San Francisco Realtors serving the Castro, Eureka Valley, Dolores Heights and the wider Bay Area since 2002. Data covers the Eureka Valley/Dolores Heights MLS district. Data from SFARMLS. InfoSparks © 2026 ShowingTime Plus, LLC. Danielle Lazier, DRE #01340326.
