What the counts are telling us as they come in, and how far they move our night forecast. Newest first; this page updates itself every 20 seconds. The numbers behind every post are on the live page.
That's us for the night. Thank you for following along.
The big picture: the Democrats have won both the House and the Senate. We called both at 8:37 pm ET. The networks called the House at around 10:30 pm ET and the Senate between about 11:25 and 11:45 pm ET, so we were roughly two and three hours ahead.
How our forecasts did. Before a single vote was counted, our published forecast gave the Democrats an 83% chance of the Senate, with about 53 seats, and a 90% chance of the House, with about 245. As the night ends, our night forecast has them on about 54 Senate seats and about 257 House seats. Its final reading is that the Democrats ran about 1 point ahead of the polls nationally, give or take under a point. Early in the evening, on the first counts, it had them as much as 4 points ahead. That fell back as Florida and the late Election Day votes came in, which is just what early counts do.
What others have confirmed so far matches our calls: Talarico in Texas, Sherrod Brown in Ohio, Roy Cooper in North Carolina, Jon Ossoff holding Georgia, and a Democratic gain in South Carolina that we hadn't seen coming.
What didn't go so well. We were too cautious on Texas for half an hour, and took down two posts that leaned on a bad estimate of the votes still to count. Florida moved its results to a new website this year and our reader needed fixing mid-evening. Some of our speculative race calls came and went, including Dan Osborn in Nebraska and Mary Peltola in Alaska, which are still being counted. We'll go through every race against what we said once the counting is done, the misses included.
Alaska, Nebraska, Maine and a fair few House seats are still to be settled, and some will take days. Our results pages keep updating as the states count. Goodnight.
It's official, or as official as election night gets: the networks have now projected that the Democrats will control the Senate as well as the House. NBC News, CNN, CBS and CNBC are all reporting it.
We called both chambers at 8:37 pm ET. The networks' Senate projections came within the last half hour, nearly three hours later, and their House projections nearly two hours after ours.
That's not luck. Our night forecast reads every count the states publish, compares it like with like against the same places in 2024, and asks one question of every piece of evidence: how far off are the polls? By 8:30 pm ET the answer was clear enough, about 2 to 4 points in the Democrats' favour with the uncertainty narrowing, that there was no realistic path left for the Republicans in either chamber.
It wasn't all smooth. Our Texas commentary wobbled for half an hour, our Florida reader needed fixing mid-evening, and some speculative race calls have come and gone. We'll go through all of that properly, and how every race compares with what we said, once the counting is done.
For now: Democrats win Congress. We'll keep going for a little while longer as the West Coast counts.
Two of our early speculative calls have now been confirmed elsewhere. CBS is reporting Democratic Senate wins in Texas and Ohio, and CNBC has Talarico projected to win Texas. We had Sherrod Brown as a speculative call in Ohio at about 8 pm ET, and Talarico not long after.
CNN's coverage says the Democrats have retaken the House and are "closing in" on the Senate. No network has called the Senate yet, as far as we can see. We did at 8:37 pm ET, and nothing since has shaken it: our night forecast still has the Democrats at 99.8%, with about 54 seats.
Two of our own speculative calls have gone the other way in the last few minutes, and it's only fair to say so. Nebraska, where Dan Osborn, the independent, is back down to 75% on our forecast, and Alaska, where Mary Peltola is at 83%. Both are still leaning their way, just not as strongly as earlier. That's the nature of calls made on partial counts.
Polls have closed on the West Coast: California, Oregon, Washington and Idaho.
The networks have now caught up on the House. CBC is reporting that the major US networks project the Democrats will take it, and CNN has made the same projection. On the Senate, MS NOW's live coverage still calls control "undecided". We called both at 8:37 pm ET.
One more Senate seat to add to the Democrats' column, if the reports hold: CNBC's coverage says the Democrats have flipped South Carolina. That was a seat we had as likely Republican this morning, and one where we haven't been able to read the state's own results feed tonight, so we'll take it from them.
Our night forecast: Democrats 99.9% for the Senate, about 54 to 55 seats, and about 261 House seats. Across 120 pieces of evidence they're running about 1.7 points ahead of the polls.
A small moment of pride. We called both the House and the Senate for the Democrats at 8:37 pm ET, on our own night forecast, built from the states' own counts.
As far as we can see, nobody else had called either chamber at that point. CNBC's coverage now has the Democrats projected to win the House, nearly two hours after our call. On the Senate, the networks are still saying "poised", while we called it more than two hours ago.
We'll be honest about the bumps along the way: we were too cautious on Texas for half an hour, and some of our speculative race calls will turn out wrong. But on the big question of the night, who controls Congress, we were first and we were right.
The House has now been projected for the Democrats elsewhere: CNBC's live coverage is reporting that the Democrats are projected to win the House and are "poised" to take the Senate.
We called both chambers at about 8:40 pm ET, nearly two hours ago, on our night forecast. The Senate call still stands on our numbers: about 54 seats, with Texas, North Carolina and Ohio among the gains, and Talarico's Texas win now confirmed by others too.
Our night forecast reads the Democrats as running about 1.6 points ahead of the polls nationally, give or take a point. That's a solid night for them, but not the landslide the very first counts suggested.
We've been watching Talarico's lead shrink and calling Texas "a genuine race". In fairness to the people who count votes for a living: it has already been called. Local 10 and The 19th are both reporting that Talarico has won, with the first reports a little after 9 pm ET.
The latest figures we can see, from KPRC Click2Houston with 57% of precincts in, back that up better than our last two posts suggested: Talarico 3,440,676, Paxton 2,628,183, a lead of about 812,000. The most recent batch barely moved it.
Where we went wrong: our sense of how many votes were still to come rested on an early estimate of about 10.4 million in total. The people who called it have far better turnout models than that one figure. If the total ends lower, there simply aren't enough votes left for Paxton, even at the margins he was winning batches by an hour ago.
So: Talarico wins Texas. Gina Hinojosa still leads Greg Abbott by about 530,000 in the governor's race, and we'll keep an eye on that one.
Texas is turning into the story of the night. On KPRC Click2Houston's figures, with just over half of precincts reporting, James Talarico has 3.29 million votes to Ken Paxton's 2.17 million: 59% to 39%, a lead of about 1.1 million.
The governor's race has followed it. Gina Hinojosa leads Greg Abbott by about 870,000 votes, 57% to 42%, and our night forecast now makes her a speculative call at 98%. Before tonight we gave her a 37% chance against a three-term governor.
The usual caution applies: what's left is weighted towards Election Day and rural votes, which lean Republican. But for Paxton to catch up he would need to win the rest by something like 20 points, and nothing we've seen in the county results tonight suggests anything close to that.
Nationally, with 114 pieces of evidence now in, the Democrats are running about 2 points ahead of the polls, give or take a point. Our night forecast: Senate 99.5% for the Democrats, about 55 seats; House certain, about 264. Still no network call on either chamber that we can find.
Nebraska is the night's surprise so far. Dan Osborn, the independent who came within about 7 points of Deb Fischer two years ago, is now a speculative call on our night forecast: 88%, up from 43% before the votes. Our forecast also has the Democrat ahead in the governor's race there. Nebraska counts slowly and its early figures can mislead, so we'll be watching this one closely. He has said he won't caucus with either party, so a win for him wouldn't change tonight's control arithmetic, but it would make the new Senate even harder for the Republicans.
Elsewhere the news outlets are catching up with seats we'd already marked down: Jon Ossoff holding Georgia, Roy Cooper taking North Carolina. CNBC now has Texas "leaning" Democratic. Still no network call on either chamber that we can find.
Texas, on KPRC's figures with 42% of precincts in: Talarico 2.52 million, Paxton 1.75 million, 58% to 40%. His lead is now about 770,000 votes, still growing. Hinojosa leads Abbott 56% to 43%.
Our night forecast: Democrats 99% for the Senate, about 54 seats; certain of the House, about 260.
Twenty minutes on from our call, we're still out on our own. CNBC's live coverage has Roy Cooper winning North Carolina's Senate seat, which we called earlier, but only has House control "leaning" Democratic. Nobody else has called either chamber yet as far as we can see.
Our numbers have come back to earth a little as Florida's own count has fed in, and that's healthy. With 94 pieces of evidence, the Democrats are now running about 1.5 points ahead of the polls nationally, give or take a point and a bit, down from nearly 4 earlier. Florida was always going to pull it back: the count there is close to what we expected.
It doesn't change the call. Our night forecast still has the Democrats at 99% for the Senate, with about 54 seats, and certain of the House, with about 260.
Texas keeps holding up for them. On KPRC's figures with 38% of precincts in, Talarico leads Paxton by about 715,000 votes, 58% to 40%, and Hinojosa leads Abbott 56% to 43%. Talarico's lead in raw votes is still growing as the count goes on.
The Texas lead is shrinking, as we said it would, but slowly. On the figures KPRC Click2Houston is showing, with 29% of precincts reporting, James Talarico has 1.80 million votes to Ken Paxton's 1.22 million: 59% to 40%. An hour ago it was 67-32 on the first fifth of the count.
The governor's race is running the same way: Gina Hinojosa 56%, Greg Abbott 43%, with 30% of precincts in. Abbott has won three times, by double digits each time.
About 3 million votes are counted out of roughly 10 million expected, so there's a lot still to come, much of it Election Day voting, which leans Republican, and the rural counties. Expect both races to tighten further. To catch up, Paxton would need to win the remaining votes by about 8 points. In the county figures we've seen, Republicans are running far behind Trump's 2024 margins even in their strongholds, so that looks a stretch.
We're calling it: the Democrats will win control of both the House and the Senate.
Our night forecast, which updates with every count from the states' own results sites, now gives the Democrats more than a 99.9% chance in both chambers. In the Senate our best estimate is about 56 Democratic seats; they need 51, because Vice-President Vance breaks a 50-50 tie. In the House it's about 280 seats; they need 218.
What's driving it is that the counts are running ahead of the polls almost everywhere we can compare like with like. Across 61 pieces of evidence, the Democrats are doing about 4 points better than the polls had them, give or take 2. In Georgia, Keisha Lance Bottoms and Jon Ossoff are 9 to 12 points ahead of where we expected. In North Carolina, Roy Cooper's Senate seat is already a firm call for us: he leads by 34 points with 61% counted, and on the like-for-like comparison he's about 12 points ahead of our forecast. Texas, Ohio, Michigan, Maine, Kansas and Iowa are all leaning the same way.
The Senate maths: the Democrats started the night with 47 seats and needed to gain four. North Carolina is in the bag on our count. Ohio (Sherrod Brown), Maine (Troy Jackson), Kansas, Iowa and Texas are all speculative calls for them on our model, each at 88% or more. They only need three of those five.
Plenty of individual races are still being counted and some of our speculative race calls will be wrong. This is our model's call, not an official result, and the networks haven't called either chamber yet. But there's now no realistic path for the Republicans to hold either.
Our night forecast now gives the Democrats more than a 99% chance of the Senate and is all but certain about the House. Across 43 pieces of evidence, the Democrats are running about 2.5 points ahead of the polls nationally, give or take 2.
Michigan has just joined in. On a tiny early count, compared like with like against 2024, the Democrats are well ahead of where we expected in the Senate race, and two House seats there, the 4th and the 10th, have become speculative calls for them.
A confession on Florida. Florida moved its election-night results to a new website this year and our reader hasn't caught up yet, so the Florida races on our night page are still moving on evidence from other states, not on Florida's own count. That's why our page currently has Leela Gray narrowly favoured in Florida's 13th, when the actual count, 99% in, has Anna Paulina Luna ahead by about 2,900 votes. Trust the count. The fix is ready and should be live shortly; see our Florida posts below for the real numbers.
Florida's House districts are nearly all counted, so here's a scorecard against our forecast.
The headline: we were close. Across the districts that are 90% or more counted, the Democrats are running less than a point better than we expected. The big swing from 2024 that everyone's talking about, Miami-Dade included, was mostly already in our numbers.
The seats to watch:
Florida's 13th: Anna Paulina Luna looks to have held on. With 99% counted she leads Leela Gray by about 2,900 votes, under a point. We had her winning by nearly 4.
Florida's 7th: a genuine coin toss. Ryan Elijah leads Bale Dalton by about 500 votes out of 310,000 with 96% in. We had Elijah by 1.
Florida's 9th: Darren Soto is just ahead, by about 1,800 votes, in a seat we had leaning Republican.
Florida's 2nd, around Tallahassee: Amanda Green leads by 5 points with 80% counted, in a seat we had the Republicans winning by 7. The rest of this district is rural, though, so that lead could shrink a lot.
Florida's 22nd: Pia Dandiya leads by 4 with 95% in, in a seat we had as a toss-up.
Going the other way, the Republicans are running a few points better than we expected in the 11th, 17th and 19th, all on the Gulf coast and in the Villages area. Florida is a mixed picture rather than a wave.
We've set the Texas county figures reported so far against two things: Trump's 2024 margin, and what our forecast expected in each county. These are counties that are 55 to 80% counted, so this is a better guide than the statewide 67-32.
The swing from 2024 is huge everywhere, about 20 points even in deep-red rural counties. Randall County, around Amarillo, went for Trump by 60 points; Paxton leads there by 40. Grayson, Rockwall, Rusk: all much the same, Paxton winning big but about 20 points down on Trump.
In those rural counties Talarico is running only 1 to 3 points better than our forecast expected. In the suburbs and cities it's 8 to 10 points better: Dallas, Williamson outside Austin, Brazoria and Fort Bend near Houston, and Denton. Brazoria is the eye-catcher: Trump won it by 20, Talarico leads by 7.
Some of the suburban lead will melt as Election Day votes are counted, since early voters lean Democratic. But even allowing for that, this looks like Talarico running ahead of our forecast, not behind it. We had him winning by 4 points. These counties suggest more.
The first statewide Texas numbers are in, and they look extraordinary: James Talarico 66.8%, Ken Paxton 31.6%, on just over 2 million votes. That's about a fifth of the votes expected.
Hold the champagne. Texas's first numbers are mostly early votes from the big cities, Houston, Dallas, Austin and San Antonio, which count fast and lean heavily Democratic. The rural counties and the Election Day vote come later, and they lean Republican. Polls in the far west of the state, around El Paso, are still open for another 40-odd minutes. A 35-point lead now could easily shrink to a few points by the end.
What does matter is the scale. The expected total is more than 10 million votes, nearer a presidential year than a midterm. Combined with Denton County, where Talarico is running far ahead of Trump's 2024 margin, it points the same way as our night forecast, which has him as an 88% favourite. We'll know a lot more when the suburban Election Day votes come in.
Texas next. Steve Kornacki on NBC has Denton County, the big suburban county north of Dallas, with about 70% counted: James Talarico 53.7%, Ken Paxton 44.9%. Trump won Denton by 13 points in 2024. If that holds, it's a swing of more than 20 points in exactly the kind of fast-growing suburb Republicans have relied on in Texas.
One county isn't Texas, and the Dallas and Houston suburbs tend to count their early votes first, which flatters the Democrats early on. But it fits with what our night forecast is already saying: we have Talarico as an 88% favourite, up from 70% this morning.
And in Georgia, NBC has Keisha Lance Bottoms on 57% with nearly half the vote counted, still too early for them to call. On the like-for-like comparison we use, she's about 12 points ahead of where we expected.
Florida is looking much closer than anyone expected.
On the state's own figures, with about two thirds of precincts reporting and most of the early and mail vote in, Ashley Moody leads Angie Nixon for the Senate by about 38,000 votes out of more than 7 million, roughly half a point. In the governor's race David Jolly is actually ahead of Byron Donalds, by about 90,000 votes. We had both races as Republican holds this morning.
The eye-opener is Miami-Dade. Steve Kornacki on NBC has been showing Nixon ahead there by about 7 points with 92% counted. Trump won Miami-Dade by more than 11 points two years ago. That's a swing of nearly 20 points in the county that did most to turn Florida red.
Closer to home for us, Florida's 13th, Anna Paulina Luna's seat around St Petersburg, is on a knife edge: Luna 149,184, Leela Gray 147,838, with 188 of 209 precincts in. We had Luna winning by nearly 4 points.
The Election Day vote still to come in Florida usually leans Republican, so Moody and Luna are entitled to feel a bit better than these numbers suggest. But this isn't the Florida of 2024.
Polls have closed across a big chunk of the country now, including Pennsylvania, Michigan, New Hampshire, Maine and the rest of Florida.
Our night forecast has made a few more speculative calls. The eye-catchers are on the Senate side: Sherrod Brown in Ohio's special election (97%, up from 69% before the votes) and James Talarico in Texas (88%, from 70%). Neither has much counted yet. They're riding on the evidence from Georgia and elsewhere that the Democrats are running ahead of the polls, so treat them as a strong hunch, not a result.
It hasn't all gone the Democrats' way. Phil Scott looks set to hold the Vermont governorship for the Republicans, and Kelly Ayotte is now an 88% favourite in New Hampshire. And our call for the Democrat in Georgia's 12th district, a seat we had at 20% this morning, rests on an early count that we expect to narrow.
Overall: Democrats 98% for the Senate, over 99% for the House, running about 1.5 points ahead of the polls nationally. Florida's count should start any minute.
Georgia keeps pulling the same way. With over a third counted, and comparing like with like, Bottoms is now about 12 points ahead of where we expected in the governor's race and Ossoff about 9 ahead in the Senate. Put together with everything else, the counts so far say the Democrats are running about 2 points ahead of the polls nationally, give or take 3.
That's enough to push our night forecast further: Democrats 97% to win the Senate and over 99% for the House.
One thing to be clear about. Our night forecast now has Democrats narrowly ahead in some races where not a single vote has been counted: the Florida governor's race (David Jolly), Texas governor (Gina Hinojosa) and, in Nebraska, the independent Dan Osborn. That's inference, not counting. If the polls are 2 points off nationally, those races tip. If Georgia turns out to be a Georgia story rather than a national one, they tip back.
Florida's results start at 8 pm ET, when the panhandle closes, along with Pennsylvania, Michigan, New Hampshire and Maine.
Now we're getting somewhere. A quarter of Georgia's vote is counted, and for the first time we can compare like with like: early votes against early votes, mail against mail, county by county with 2024.
On that fair comparison, Keisha Lance Bottoms is running about 11 points better than we expected in the governor's race against Rick Jackson, and Jon Ossoff about 8 points better against Mike Collins. That's no longer count-order noise. Georgia is now the main thing moving our forecast, and the night forecast has Bottoms all but certain.
Not everything is going one way. Virginia's Senate race, on just 3% counted but compared like with like, has the Republicans about 7 points better than we'd expect. And in Vermont, the early count has Phil Scott, the popular Republican governor, ahead again, which we'd half expected.
The Indiana and Georgia district numbers that drove the first jump have faded in importance now that better evidence is in. Overall: Democrats 94% for the Senate, 99% for the House, running about a point ahead of the polls nationally. North Carolina and Ohio closed at 7:30, and their counts are next.
The first real counts are in and our night forecast has jumped: Democrats now 94% to win the Senate (from 84%) and 98% for the House (from 91%).
Before anyone gets excited, a word of caution. Most of that move comes from tiny, early counts. The single biggest driver is Indiana's 2nd district, where with about 4% counted the Democrat is running 29 points ahead of where we expected. In Georgia's 9th, with 30% in, it's 55 points better than expected. Those numbers will not hold. States count their early and mail ballots first, and this year those lean heavily Democratic. Going the other way, Virginia's 4th, a safe Democratic seat, is running 52 points worse for them on 3% counted.
Our model knows early counts are skewed and gives each one only a little weight, but there are now 21 of them and they mostly point the same way. Taken together, they say the Democrats are running about 1.3 points ahead of the polls nationally, give or take more than 3 points. That's the honest read: a slight lean, plenty of room either way.
It has also thrown up our first speculative calls, which are exactly that: speculative. We now have Troy Jackson as a strong favourite to unseat Susan Collins in Maine, and Amy Acton for Ohio governor. Expect some of these to wobble as the Election Day votes come in.
That's the big one: polls have closed in Georgia, Virginia, South Carolina, Vermont and most of Florida, along with the rest of Indiana and Kentucky.
We're reading Georgia, Virginia and Indiana straight from the states' own results sites, and they're all up and running. They just haven't got many votes on them yet. Georgia and Virginia usually post their early and mail ballots first, so the first numbers tend to flatter the Democrats. We compare them with the same kind of votes in the same counties in 2024, so we shouldn't be fooled by that.
What we're watching first: Jon Ossoff's Senate seat and the governor's race in Georgia, where we have the Democrat at 62%, and the Virginia House seats. Virginia is the cleanest early read on the suburbs anywhere in the country tonight.
Still 84% for the Democrats in the Senate and 91% in the House while we wait.
Still no counted votes in our feeds, so the forecast hasn't moved: Democrats 84% for the Senate, 91% for the House.
The first exit poll numbers are out, though. CNN is reporting that 71% of voters in its exit poll think Congress isn't doing enough to check President Trump. Exit polls on the night are rough and get adjusted as the real count comes in, so we don't put them into the model. But a number like that is the kind of mood the polls had been picking up all autumn, and it's the backdrop the Republicans were hoping to escape.
The big test starts in 40 minutes, when polls close in Georgia, Virginia, South Carolina, Vermont and most of Florida. Virginia and Georgia count quickly and break their results down by how people voted, so they'll be the first states where we can compare like with like.
Polls have now closed across most of Indiana and Kentucky. The western ends of both states, which run on Central time, vote for another hour.
Nothing counted has reached us yet. Indiana's results come straight from the state's own election site, and we'll start reading them as soon as the first counties report. Kentucky is harder: the state's results site turns away automated readers like ours, so for Kentucky's 6th district we'll be relying on what the counties and other outlets report, and it won't move our forecast.
For now we're where we started: Democrats 84% to win the Senate, 91% for the House. The big wave of closings is at 7 pm ET: Georgia, Virginia, South Carolina, Vermont and most of Florida.
Before any votes are counted, the early vote is the only real evidence we have, and so far it's mildly encouraging for the Democrats.
In Pennsylvania, registered Democrats' lead among the ballots already cast is 17 points bigger than it was on the same day in 2024. In North Carolina it's 13 points bigger, and in Maine about 10. Given how the national mood has moved since 2024, we'd have expected something like 9 points. So Pennsylvania and North Carolina are running ahead of that, and Maine is about where we thought.
Don't get carried away. This is party registration, not votes: plenty of registered Republicans vote Democratic and vice versa, and who votes early changes from one election to the next. We treat it as very noisy evidence, which is why it has only nudged our Senate chance from 83% to 84%.
The first real numbers come from Indiana and Kentucky, where most polls close at 6 pm ET. The one we're watching is Kentucky's 6th district, which the Wall Street Journal picked out as the earliest House race worth reading. Trump won it by 15 points in 2024, and we have the Democrat as a 62% favourite. If that holds up early, it's a good sign for the Democrats' night.
Evening all. The first polls close in parts of Indiana and Kentucky at 6 pm ET, with Georgia, Virginia, South Carolina and Vermont following at 7 pm. We'll be here through the night.
Where we start: an 83% chance the Democrats win the Senate, about 53 seats on average, and 90% for the House.
The Economist sees the Senate very differently. Its model has it as close to a coin flip, 52% for the Democrats, with a middle outcome of 51 seats to 49, and it says itself that it's now a substantial outlier: the betting markets and the other models all have the Democrats as clear favourites.
The reason is a handful of states Donald Trump won by double digits in 2024. On blue and purple states The Economist is roughly where everyone else is (North Carolina, Maine and Michigan all lean Democratic). But in the deep-red ones it expects the polls to be wrong in the Republicans' favour, and the Democrats need to win at least two of them.
That's where we part company. In Texas, Ohio, Kansas and Alaska our forecast has the Democrats (Talarico, Brown, Hamilton and Peltola) three to four points ahead. The Economist has Texas and Ohio dead level and puts Iowa, South Carolina, Nebraska and Florida in the Republican column.
It's a fair worry. Polls in red states have flattered Democrats before, and our model allows for the polls missing the same way everywhere: it's a big part of why about one of our simulations in six still ends with a Republican Senate. We just don't build in a red-state miss before we've seen one.
So Texas and Ohio are the ones to watch. If Talarico and Brown run behind their polls when the counts come in, The Economist will look very clever. Our night forecast moves with every count; the early vote has nudged the Senate up a fraction, to 84%. Nothing to read into that yet.
