Democrats are the clear favorite to take the House in November, and Kalshi's traders and our price-blind AI panel agree on that. They part ways on the Senate. The market makes Democrats a 69-cent favorite there. The panel, working from the actual list of seats in play, puts it near one in five. That single gap decides which of the four possible Novembers is the favorite. The market's top pick is a full Democratic sweep at 68 cents. The panel's favorite is a split Congress, Democratic House and Republican Senate, which it prices at 55 percent against a market price of 23 cents.
The seat-by-seat board, the Senate math behind the call, and the individual races the panel has scored are below. Prices are Kalshi's as of October 23, 2026. The panel's estimates were generated August 8 and have not been re-run since.
The Market
| Venue | Kalshi (where prediction markets are legal) |
| The Event | "2026 Midterms: Congress Balance of Power" (ticker KXBALANCEPOWERCOMBO-27FEB), four mutually exclusive outcomes, one contract each |
| Settles | On which party controls each chamber as of February 1, 2027, decided by the November 3, 2026 elections |
| Prices As Of | October 23, 2026: D-House/D-Senate 68¢ · D-House/R-Senate 23¢ · R-House/R-Senate 8.7¢ · R-House/D-Senate 0.6¢ (cost to buy Yes) |
This is a neutral, data-driven analysis. It presents both parties' paths identically, advocates for neither, and every electoral fact in it comes from the sources named on the page.
The Board
Six models returned verdicts on this run; the GPT and Gemini seats were unavailable. None of them saw Kalshi's prices. Each priced all four outcomes in one pass, so no model could favor a Democratic House in one row and contradict itself in another.
| Outcome | Market | Fable | Opus | Sonnet | GLM | Kimi | DeepSeek | Blend | Edge |
|---|---|---|---|---|---|---|---|---|---|
| D House / R Senate (Split) | 23¢ | 58 | 55 | 61 | 44 | 58 | 55 | 55.2% | +32.2 |
| D House / D Senate (Sweep) | 68¢ | 20 | 21 | 14 | 8 | 22 | 20 | 17.5% | −50.5 |
| R House / R Senate (R Hold) | 8.7¢ | 20 | 21 | 21 | 41 | 18 | 20 | 23.5% | +14.8 |
| R House / D Senate (Reverse) | 0.6¢ | 2 | 3 | 4 | 7 | 2 | 5 | 3.8% | +3.2 |
Read by chamber, Kalshi's separate House and Senate markets have Democrats taking the House at 91.5 cents, and the panel, from its August 8 run, has it at 72.7 percent. On the Senate the market sits at 69 cents and the panel at 21.3.
The panel's estimates were generated August 8, 2026, price-blind, from the data card described below; market prices were re-fetched October 23, 2026. Models are frequently wrong, and the market price reflects real traders' money. Our full graded record, including the events we have lost to the market, is public. These are market prices and estimates, not predictions of fact and not financial advice.
Why Both Sides Expect A Democratic House
The president's party has lost House seats in 18 of the 20 postwar midterms, an average of 25.6. The two exceptions, 1998 and 2002, each needed a dominant national event to flip the usual referendum dynamic, and nothing like that is on the board now. The panel's last firm polling read, from late June, had the generic ballot at Democrats +5.0 and presidential approval at a net of −17.8. Midterm losses run worse for presidents that far under water. The panel is more cautious than the market, 72.7 percent against 91.5 cents, because of the weeks that were still to run when it priced this, but the two readings point the same way. The gap widened this week: the House rung had closed between 90.0 and 90.4 every day since October 7, then closed 90.8, 91.3 and 91.4 while the panel's number stood still.
Why The Panel Fades A Democratic Senate
Democrats need a net gain of four seats, and only four Republican-held seats were rated competitive on the panel's August 8 read. Every other Republican seat, including the open ones in Alabama, Iowa, Kentucky, Montana and Wyoming, sits in a state the party wins by double digits. Three Democratic seats are competitive too, and Michigan is a true tossup.
| Seat | Held by | Status | Rating (August 8) |
|---|---|---|---|
| North Carolina | Republican | OPEN (Tillis retiring) | Lean D flip, the best target on the board |
| Maine | Republican | Collins, incumbent | Tossup |
| Alaska | Republican | Sullivan, incumbent | Tossup |
| Ohio (Special) | Republican | Husted, appointed | Tossup |
| Michigan | Democrat | OPEN (Peters retiring) | Tossup, a live Republican flip chance |
| Georgia | Democrat | Ossoff, incumbent | Lean D |
| New Hampshire | Democrat | OPEN (Shaheen retiring) | Lean D |
So a Democratic Senate means winning essentially every competitive Republican seat in the country while holding all three defenses. Lose Michigan and the requirement becomes a fifth pickup from the safe-Republican tier. A House wave lifts a party's candidates a few points everywhere, which flips dozens of close districts, but a few points does not turn Alaska or a Maine incumbent into a safe seat. In 2018 Democrats gained 41 House seats and still lost two in the Senate, because the map decided the upper chamber. Every model on the panel made the split its single most likely outcome for that reason.
The Case For A Republican Hold
The market prices a Republican hold at 8.7 cents and the panel blends it at 23.5. Three of the six models opened with the Republican hold as their most likely outcome, reasoning from midterm gravity and the map, then moved to the split after weighing the House base rate. GLM kept a unified Republican Congress as a loud second at 41 percent, arguing that low presidential approval could prove more decisive than the generic ballot. If the June generic-ballot lead has faded since the panel's last polling read, this is the outcome that benefits. The reverse split, a Republican House with a Democratic Senate, is nobody's base case at 0.6 cents; the panel prices it like a lottery ticket at 3.8.
What Moved Since October 20
The chamber moved, five cents in three sessions, as much as any seat under it. Kalshi's separate Senate market is 69 cents to buy, from 64 at this page's last update. It closed at 65 on Tuesday, 66 on Wednesday and 69 on Thursday, its highest close on this page's daily record, which runs from September 25 and had peaked at 65 on October 5. Kalshi's news desk called 66 "a new high" in a Wednesday post, noting the contract "traded around 47% at the beginning of September," and it closed three cents above that new high on Thursday. The sweep followed to 68, from 64, after a 69 close on Thursday. The split the panel favors is 23 cents, from 27, after closing at 25 on Tuesday, 27 on Wednesday and 23 on Thursday, under the 27 close of September 30 that had been its floor on this page's daily record. The House market is 91.5, from 90: it had closed between 90.0 and 90.4 every day from October 7 through Monday, then 90.8, 91.3 and 91.4. The Republican hold is 8.7 cents to buy, from 9.6, and the reverse split is 0.6, from 0.8.
Two seats carried it, and they stood in opposite relation to the polls.
Ohio moved on a poll, two of its five cents in the hour after it printed. Sherrod Brown is 70 cents to buy, from 65, after closing at 69 on Wednesday and 70 on Thursday, the two highest closes the seat has on this page's record. Marist's Ohio poll, released Tuesday from October 15 to 18 fieldwork, had Brown 48 to Jon Husted's 44 among likely voters, and the price did nothing with it: Brown closed at 65 on Monday and 65 on Tuesday. Fox News published its Ohio poll at 6 p.m. Wednesday, Brown 52 to 47 among 1,592 likely voters, with Husted's favorability 13 points underwater, 42 percent favorable to 55 unfavorable. The seat had already taken two cents in the hour ending at 2 p.m., before the Fox poll was public, 65 to 67 on 14,035 contracts, a tick I cannot attribute; the hour ending at 7 p.m. took the next two, 67 to 69, on 49,335 contracts, the seat's busiest hour of the week. Fox's Dana Blanton wrote that "about 2½ times as many voters say their financial situation is worse than two years ago rather than better (46% vs. 19%)." Two polls a day apart pointed the same way; the market skipped the first and took the second. Two updates ago this seat jumped five cents on a Bowling Green State University poll. It has now done it twice.
Texas moved through a poll. James Talarico is 66 cents to buy, from 63, and Ken Paxton is 34, from 38. Talarico closed at 64 on Tuesday and 68 on Wednesday, his highest close on this page's record, above the 67s of October 6 and 7, then 67 on Thursday; the ask eased to 66 overnight. The Wednesday move started in the hour ending at 2 p.m. ET, when 220,808 contracts went through on the Talarico side, the seat's busiest hour of the week, and the close went from 62 to 65; it was 68 by the hour ending at 7 p.m. What was public by then: Harris County's record first day of early voting, reported by the Texas Tribune on Tuesday alongside records in Dallas, Montgomery, Ellis, Bell and McLennan counties, with Republican consultant Ross Hunt cautioning that "usually the aggregate level of turnout or percentage level turnout really doesn't bear directly on the outcome"; and Michael Bloomberg's $2.5 million to the super PAC backing Talarico, the only Senate race he has funded this cycle per the Washington Post's Tuesday report, alongside the Times' Tuesday report on his $80 million. Neil Bush's Facebook post that he had voted for Talarico, "motivated by two things: deep concern about Ken Paxton's record and genuine admiration for the character and principles I see in James Talarico," surfaced the same day, hours before Talarico's University of Houston speech per Houston Public Media.
What came after the move: Fox News at 6 p.m. Wednesday had the race tied at 49 among 1,398 likely voters, from Talarico 51 to 49 in September, writing that "the shift to a tie reflects Paxton consolidating support among conservatives, rural voters, suburban women and voters who backed Cornyn in the primary," and Emerson on Thursday had Talarico 48 to 46 with five percent undecided, Spencer Kimball writing that if the Libertarian's support "melts between now and Election Day, Paxton will likely see his vote share increase." The price closed at 68 the night of the tie poll and 67 the day of the two-point one. I cannot put a headline on the 2 p.m. hour. I can say that the public record it had to work with was a turnout number, a check and an endorsement, and that two polls pulling the other way cost it a cent. Harris County's third day was 93,417 in person, 327,691 early votes through Wednesday on the county clerk's unofficial count, and the Paxton side traded 1,675,872 contracts in the 24 hours before this pull, more than any market on this page, the House market included.
Kansas got the rescue I said was scheduled, and the seat ended the week where it started. Roger Marshall is 62 cents to buy, from 61, and Adam Hamilton is 39, from 40; Marshall closed at 63 on Monday, 61 on Tuesday, 62 on Wednesday and 61 on Thursday. President Trump's rally in Park City on Monday evening, the night before the last pull, drew "more than 6,000 people" per KCUR; ABC News reported that Trump called Hamilton a "bad person" and a "disgrace" who is "pretending to be a moderate," and that Hamilton's spokesperson Tyson Brody answered, "Roger Marshall's campaign is sinking, and Washington is panicking." Wednesday evening brought the last scheduled debate, where Marshall said, "Control of the Senate goes through Kansas," and, "This guy is funded by Chuck Schumer. He'll be beholden to Chuck Schumer forever," and Hamilton said, "(Marshall) says he speaks up for us, but the president doesn't seem to listen." KCUR's reference poll is still the New York Times/Siena survey fielded September 22 to 30, at 45 to 45 with six percent undecided. Marshall's close fell two cents the session after the rally; three sessions and a debate later he is back where he closed on Tuesday, and a cent higher on the ask.
Maine traded more than it moved. Troy Jackson is 65 cents to buy, from 64, after a 65 close on Thursday, and Susan Collins is 37, where she was. The news Kalshi's desk attached to the price on Thursday morning was old by this page's clock: Joe Manchin's endorsement of Collins, "She's not Donald Trump's senator. She's not the Republican senator. She's Maine's senator," reported by the Times on Monday the 19th, the day before the last update, and a Washington Post-Schar School poll with Jackson 50 to 46 that was fielded September 30 to October 5. What was new was the volume: the Collins side traded 402,559 contracts on Thursday, the most either Maine side has done in a day on this page's record, for a close one cent from Wednesday's.
The rest of the board moved a cent or less. Abdul El-Sayed is 69 in Michigan, from 68, and closed at 69 every day since Monday, including Wednesday's final debate, where per the desk Mike Rogers said, "You want to pay $8 a gallon for gas? Let Iran get a nuclear weapon," and El-Sayed said, "He's got a war for everything." Iowa went two cents toward the Republican without Ashley Hinson's price moving: she is 58, where she was, and Josh Turek is 42, from 44, the two asks summing to 100 now where they summed to 102 at the last update. The Democratic line in Alaska, Mary Peltola's, is 76 to buy, from 75, after 76 closes on Wednesday and Thursday, with Data for Progress putting her at 47 to Dan Sullivan's 44 on first choices and 51 to 49 in a simulated final round among 694 likely voters polled October 8 to 16. Roy Cooper is 96.5 in North Carolina, from 96. Chris Pappas is 90 in New Hampshire, unchanged.
Set it against last update's finding, that the board was trading the campaigns and its own order flow rather than the public polling. Texas confirmed it: the seat moved through a poll that said tie and one that said two points. Ohio half-refuted it: the market skipped the first ballot line and bought the second, inside the hour. The chamber moved five cents on the pair. What the two seats share is the direction of the rest of the evidence, Ohio's favorability gap, Texas's turnout and money. Kansas, where the week's non-poll evidence was a presidential rally, is the control: Marshall's close fell two cents the session after it, and the seat ended the week where it began.
The update before this one belonged to Maine, which moved seven cents without a poll: Jackson went from 57 to 64 after 27,500 contracts went through in the hour ending at 2 a.m. ET on Tuesday the 20th, in the weekend the two candidates spent relitigating his "Oh, shut up" at the Bangor debate on national television, which the Portland Press Herald reported alongside Jackson's "I'm not going to apologize for that." Kansas went the other way that week, Marshall 65 to 61, after Sabato's Crystal Ball moved the race to Toss-up on October 14 and NPR reported that both parties "have redirected tens of millions of dollars in recent weeks"; Texas got two polls within the margin, CBS News/YouGov at 51 to 49 and a Hobby School poll at 48 to 46 via Kalshi's desk, and moved a cent; and the finding was that polls arriving did nothing while seats without one moved.
The three updates before that were Ohio, the stop and the slide. Brown went from 60 to 65 on the October wave of the Bowling Green State University/YouGov poll, reported by Political Wire at Brown 48 to Jon Husted's 42, in the week Politico's tally of third-quarter reports had the Democratic campaigns in the ten competitive Senate races raising more than five times what the Republican campaigns raised. Before that, between this page's October 7 and October 10 updates, the Senate market went from 65 cents to 61 and the sweep from 65 to 60, in the week of the first Maine and Michigan debates and Talarico's absence in Texas; Kalshi's news desk put the Michigan drop after the debate at six points, from El-Sayed's 73 close of October 5 to 67, priced Collins at 44 the day after her first debate, citing a Marist poll with Jackson at 51 and Collins at 47 among registered voters, and had Marshall at 62 in Kansas on October 8, down from 82.8 on September 5, after the Cook Political Report moved the race to Toss Up and Cook's Jessica Taylor wrote, "At this point, the question isn't whether Democrats will gain seats, but how many," revising the range of potential Democratic pickups to two to six seats. By the October 13 update the slide had stopped at 62, and Talarico had won back two of the four cents he lost.
Those weeks sit on top of September's run, which is still the bigger move. The sweep, 44 cents when this page opened on August 8, closed between 46 and 48 cents every day from August 18 through September 13, then went from 48 to 59 in four trading days, September 14 through 17, when Kalshi's separate Senate market flipped to Democrats. Kalshi's news desk marked that flip at 52 to 48 on September 14, pointing to an Associated Press report in which Senate Majority Leader John Thune told South Dakota's KELO-TV, "I'm a realist. I don't ever try and sugarcoat things. I think it's a competitive environment right now," and quoting Senate Minority Leader Chuck Schumer, "We found new states – Iowa, Texas – which people a year ago weren't even paying attention to." Time's Philip Elliott, quoted by Kalshi's news desk on September 29, reported that the Senate Republicans' campaign chairman, Tim Scott, told senators the committee's priority was "saving GOP-held seats in Texas, Iowa, Maine, and Alaska," according to a person who listened to the briefing.
The panel's numbers stay frozen where August 8 left them, so the sweep gap it fades, 26.5 points when this page opened on August 8 and 46.5 at the last update, is 50.5, the widest on this page. The frozen table above says a Democratic Senate "means winning essentially every competitive Republican seat in the country while holding all three defenses." Price by price, the market now has five Republican-held seats more likely than not to go Democratic, North Carolina at 96.5, Alaska at 76, Ohio at 70, Texas at 66 and Maine at 65, with Iowa at 42 and Kansas at 39 for the Democrat, and all three defenses at 69 or better. Democrats need a net gain of four. At these prices the market is not asking them to win every competitive Republican seat; it has them winning five of the seven and losing Iowa and Kansas, one pickup more than the four they need. Iowa and Kansas were not on the panel's August competitive list either; Kansas has no row in the table above because in August it was not a race. If Texas and Ohio are where the board has them, that is the whole argument for 69 cents. If they are not, the Senate market made its new high on Ohio, which the panel's August read called a tossup, and Texas, which that read did not count as competitive at all; Cook and Sabato had it Lean R. The panel has not been re-run since August 8; its four-way call stays on the public scoreboard and gets graded against the result either way.
The Senate Races That Decide It
Each row links to that race's full model verdict.
| Race | Seat | Cook / Sabato rating (August 8) | Our verdict |
|---|---|---|---|
| North Carolina Senate | OPEN, Tillis retiring | Lean D flip | Cooper vs. Whatley |
| Maine Senate | Collins (R) incumbent | Tossup | Collins vs. Jackson |
| Georgia Senate | Ossoff (D) incumbent | Lean D | Ossoff vs. the field |
| Michigan Senate | OPEN, Peters retiring | Tossup | El-Sayed vs. Rogers |
| Minnesota Senate | OPEN, Smith retiring | Likely D | Flanagan vs. the field |
| New Hampshire Senate | OPEN, Shaheen retiring | Lean D (Sabato) | Pappas vs. Sununu |
| Iowa Senate | OPEN, Ernst retiring | Likely R | Hinson vs. Turek |
| Texas Senate | R-held (Paxton primary) | Lean R | Paxton vs. Talarico |
The chamber-level boards from the same panel: House Control 2026, 2026 Senate control and the full Balance of Power four-way breakdown. New to election contracts? Start with how election prediction markets work.
Every verdict above is stored as a graded row on our public model scoreboard, and each model's four-way call gets scored against the actual result when this market settles, win or lose.
This analysis advocates for no party and no outcome.
Prices on this page are Kalshi's book. If you also trade on Polymarket, code OS4 gets new users a $50 trading bonus on a $10 deposit — affiliate link; terms as stated by Polymarket; 18+, availability varies by state.
More on this: Brazil Election 2026: Trump's Tariff Is Boosting Lula · Israel Election 2026: Bennett Calls Netanyahu an 'Anarchist · Germany's AfD Is One Vote From Its First State Government · South Carolina Senate Runoff Odds: Graham Vs Norman, Kalshi · Kalshi Election Predictions Scorecard: Florida's Misses
