Jon Ossoff is the heavy favorite to keep Georgia's Senate seat, and Kalshi traders have quietly stopped treating his race as a fight at all. The contract on a Democratic hold sits at a 93.0¢ midpoint, through a summer in which his Republican opponent generated a steady run of damaging headlines. So we handed the race to eight AI models, each shown every verified fact and every public poll, and no market price. They agree on the winner. They would not follow the market into the 90s.
The Quick Answer
Kalshi prices a Democratic hold of the Georgia Senate seat at a 93.0¢ midpoint, making incumbent Jon Ossoff a heavy favorite over Republican Mike Collins, whose side trades near 7¢. Our price-blind panel of eight models lands on an 80% chance a Democrat holds the seat, which leaves Collins about 20%, roughly three times the chance the market gives him.
The panel sits about 13 points below the market on the Democrat. One thing changed in Ossoff's favor this summer, the models credited it, and they still stopped at 80. Why is the rest of this page.
Stay ahead of the markets.
Daily insights and expert picks on Kalshi, Polymarket, and what's moving markets.
Free forever. Unsubscribe anytime.
The case for each side of the Georgia (GA) Senate race, priced on Kalshi: Why Mike Collins (R) Will Win The Georgia Senate Race · Why Jon Ossoff (D) Will Win The Georgia Senate Race
The Market At A Glance
| Market | Which party wins the 2026 U.S. Senate race in Georgia |
| Venue | Kalshi, a CFTC-regulated event-contract exchange. 18+ only, and you must be in an eligible state; availability varies by state and can change (checked September 4, 2026) |
| Ticker | SENATEGA-26, paired contracts on the Democratic and Republican outcomes |
| Democratic side | 92.2¢ bid, 93.7¢ ask |
| Republican side | 6.6¢ bid, 7.0¢ ask |
| Open interest | About 207,000 contracts on the Democratic side and about 610,000 on the Republican side |
| Settles | On the party of whoever is sworn in for the term beginning in 2027, not on either name. A third-party winner resolves both contracts NO. A December 1 runoff would fall inside the contract's life rather than voiding it |
| Panel read | 80% Democratic win, 20% Republican |
| Prices as of | September 4, 2026 |
Why Is The Runoff Off The Table?
This is the fact that moved our models more than any poll, and the one most coverage of this race has skipped.
Georgia requires 50% plus one to win a statewide election. Fall short and the top two go to a runoff four weeks later, which this cycle falls on December 1. That rule decided both of the state's last two regularly scheduled Senate elections, Ossoff's own in 2020 and Raphael Warnock's in 2022. David Perdue actually led the first round with 49.73% to Ossoff's 47.95%, and Ossoff won the seat in the January 2021 runoff. Warnock led his first round with 49.44% and finished the job in December.
In both cycles a Libertarian on the ballot took the majority away. Shane Hazel drew 2.32% in 2020, and Chase Oliver drew 2.07% two years later. Two candidates under three points, two races completely reset.
There is no Libertarian on Georgia's ballot this year. The Libertarian Party of Georgia lost its automatic ballot line after its 2024 presidential nominee drew 0.4% of the state's vote. The law requires 1%. That left the petition route, which needed roughly 72,000 signatures by a deadline in mid-July, and the party said it could not meet it.
The Atlanta Journal-Constitution reported the direct consequence in July: the Ossoff-Collins race "will now almost certainly be settled with the Nov. 3 vote." Inside Elections cited the same change when it moved the race off Tossup on August 6.
A runoff was never automatically bad for a Georgia Democrat. They are two-for-two in these second rounds, and won the January 2021 one on lower turnout than the first round. What has gone is a second coin flip, with a fresh national money surge attached, that is now far less likely to be reached. The rule itself has not changed and the December date is still on the calendar, so the right number for it is small rather than zero. Shrinking that path is worth real money to a favorite, and it is a large part of why the models moved up.
Which leaves the first coin flip, the one the ballot change did nothing about: Georgia itself. That is the one the panel keeps pointing at and the price does not.
Why Didn't Collins's Summer Move The Price?
The Democratic contract sat near 92¢ on August 18 and its midpoint is 93.0¢ now. Across the weeks described below, it moved about a penny, which is inside the spread.
Collins won the Republican nomination on June 16 with 55.5% in a runoff against former football coach Derek Dooley, after President Trump endorsed him. Since then the documented record has accumulated steadily. The Washington Examiner, reporting CNN's reporting in July, described Collins's son-in-law David Alan Scheer II as having "shared posts supporting the white nationalist group Patriot Front, circulated Nazi imagery, called for the deportation of Muslims, and posted antisemitic messages targeting Jews." The same reporting covers a House Ethics Committee investigation of his former chief of staff, Brandon Phillips, over allegations that the office used taxpayer funds for campaign work, which Collins dismissed as a "nothing burger." At a Savannah campaign event reported on August 13, a convicted Capitol rioter asked Collins about compensation for January 6 participants, and the Atlanta Journal-Constitution recorded his answer: "when the federal government harms you, then yes, they have a liability too." His spokesman said afterward that Collins does not support a blank-check fund or automatic taxpayer-funded payouts.
What the record demonstrably produced is Republican cover for the Democrat. CBS News Atlanta reported in August that former Republican Senator Jeff Flake and two sitting Republican mayors, Scott James Matheson of Valdosta and Julie Smith of Tifton, all endorsed Ossoff. Smith's reason, in full: "I'm a Republican, but there are times when a person exceeds the party." Whether any of it moved votes, the sourcing does not show. Ossoff has his own attack lines to answer, including criticism of his Israel voting record from the publisher of the Atlanta Jewish Times and Republican advertising centered on trans athletes in school sports.
A contract already in the low 90s cannot express "this got somewhat worse for the challenger" in any way you would notice. When a price stops responding to a story, the useful question is whether the price ever left enough room for the story to matter.
What Does Each Side Have?
For Ossoff
Four public head-to-head surveys have been taken since both nominees were set on June 16, and Ossoff leads every one of them, by an average of 8.3 points.
| Pollster | Field dates | Sample | Ossoff | Collins |
|---|---|---|---|---|
| InsiderAdvantage (R) | August 16-17, 2026 | 800 LV | 50% | 43% |
| Fabrizio Ward (R)/Impact Research (D), for AARP | July 13-16, 2026 | 1,060 LV | 52% | 43% |
| Wick | June 27-30, 2026 | 1,175 LV | 47% | 43% |
| Beacon Research (D)/Shaw & Co. (R), for Fox News | June 23-27, 2026 | 1,002 RV | 56% | 43% |
The most useful row is the first one. InsiderAdvantage has a Republican-leaning house history, and it still had Ossoff at 50%, exactly on the majority threshold this whole race turns on. Six published aggregators put the race between Ossoff by 6.8 and Ossoff by 8.9 points.
The national picture helps him. Presidential approval sits near 38% approve against 58.7% disapprove, and the generic congressional ballot runs about six points Democratic.
His money is the other half. Federal Election Commission filings covering the period through June 30 show Ossoff holding $42.59 million in cash. Collins reported $2.08 million, a gap of roughly 20-to-1 heading into the expensive months.
Every handicapper on the board has the race on the Democratic side. Here is the whole board, so you can check the tiers rather than take the count:
| Rating | Raters, with the date of each rater's current rating |
|---|---|
| Likely Democratic | Decision Desk HQ (Aug 28) · The Economist (Aug 31) · FiftyPlusOne (Aug 30) · Race to the WH (Aug 30) · Sabato's Crystal Ball (Aug 26) · Silver Bulletin (Aug 21) · Split Ticket (Aug 31) |
| Lean Democratic | Cook Political Report (Aug 20) · Fox News (Aug 18) · RealClearPolitics (Aug 27) |
| Tilt Democratic | Inside Elections (Aug 6) |
For Collins
Sort that same board by tier rather than by side and the picture changes. Seven say Likely, three say Lean, one says Tilt, and not one says Safe or Solid. The professionals who watch this for a living are unanimous about the direction and unanimous that it is not finished. The five most recently dated ratings, all from the last days of August, sit at Likely Democratic, so their direction of travel is toward the market.
Georgia's own record is the core of his case. Every major statewide federal race here since 2020 has finished inside three points.
Joe Biden carried the state by 0.24 points, and Ossoff won this seat by 1.22 in the runoff. Raphael Warnock won the other seat by 2.80, and Trump took the state back by 2.20 in 2024. Nothing in the modern history of this state says a Democrat wins it comfortably, and the one clean trend line is Trump reclaiming the top of the ticket after Biden had taken it.
Then there is what nobody has measured. The newest public poll in this race was in the field on August 16 and 17. That is a gap of 18 days in an active Senate race, across the exact stretch when a nationalized contest starts spending real money.
With no third line on the ballot, the published average has to divide between Ossoff at 50.8% and Collins at 42.9%. Hand Collins every one of the unallocated points and the count still finishes Ossoff 50.8 to 49.2. So the undecideds alone do not get him there. For the win you need the polling average itself to be wrong, and Georgia's record is the case that it can be.
The market half-agrees with him on the other line of the same ballot. On Georgia Governor 2026, Kalshi prices Democrat Keisha Lance Bottoms at a 56.5¢ midpoint while pricing Ossoff at 93.0¢ directly below. Its combo market puts a Rick Jackson-for-governor, Ossoff-for-Senate split at 36.5¢, so better than a third of the money on that board expects this electorate to split its ticket in November. That is the closest thing the market offers to the argument the panel is making.
Model Verdicts
Market: a 93.0¢ midpoint on a Democratic win. Panel: 80%. The eight models sit about 13 points below the market.
Every seat scored the race without seeing a price, working from the same fetched block of race facts, polls, ratings, money and the ballot-access change. Then each read the others' anonymized reasoning and could revise. These are the revised numbers.
Every Seat's Number
| Model | Democratic win | Graded record | The one-line reason |
|---|---|---|---|
| Claude Opus | 86% | 87% on 1,621 graded calls | Ossoff leads all four post-nomination polls by about eight points, with a favorable national environment and a 20-to-1 cash edge. |
| Gemini | 84% | 86% on 6,624 graded calls | An eight-point polling average and a large cash advantage, offset by an unpolled late-summer window and Georgia's structural Republican lean. |
| Kimi | 82% | 84% on 3,119 graded calls | The post-nomination lead holds even among historically Republican-leaning pollsters, in a midterm where polls have been nearly unbiased. |
| GPT | 80% | 85% on 7,942 graded calls | The polling lead, the national midterm fundamentals and the absent Libertarian ballot line outweigh Georgia's close partisan baseline. |
| Claude Sonnet | 80% | 85% on 1,597 graded calls | A consistent lead and no runoff threat, tempered by a state Trump carried by 2.2 points in 2024. |
| GLM | 80% | 82% on 2,941 graded calls | A strong national environment plus the polling and cash leads, compounded by the Libertarian absence removing the runoff mechanism. |
| DeepSeek | 78% | 81% on 2,985 graded calls | The polling lead and financial advantage in a midterm environment favoring the out-party. |
| Grok | 78% | 84% on 5,579 graded calls | Revised up for the ballot-access change and Republican-pollster confirmation, not for peer clustering. |
| Blended verdict | 80% | Recomputed daily | seat median of eight models after the revision round (method always disclosed) |
Every seat on this panel is graded against real market settlements; the full scoreboard is public.
Model estimates generated September 4, 2026; market prices as of September 4, 2026. These are model estimates, not predictions of fact and not financial advice. Models are frequently wrong; the market price reflects real traders' money. Election markets are presented for information only, with no endorsement of any candidate or party.
Model names are trademarks of their respective owners; the verdicts are our own outputs and imply no endorsement by any model maker.
Kalshi is a CFTC-regulated event-contract exchange, not a sportsbook. Trading is 18+ and you must be located in an eligible state; availability varies by state and can change.
Every number in this piece gets graded in public once the market settles, on the full graded scoreboard.
Where The Panel Changed Its Mind
In the first pass the seats ran from 74% to 88%. After reading each other they ran from 78% to 86%.
The middle answer shifted a single point, from 79% to 80%. Five seats moved up and three moved down, and nearly every one named the argument that moved it.
The seat that started lowest moved four points, and said plainly that it was moving on evidence rather than on company.
"Prior 74% underweighted Inside Elections’ ballot-access move off Tossup and R-pollster confirmation (+4 to 78%). Strongest opposing 85–88% cases do not move further: no Safe D rating, knife-edge federal history, and the unpolled window keep a material NO path." — Grok
The largest downward move came from a seat that decided its peers were right about the state.
"Revised down from 85% toward the cluster median after weighing forecasters who emphasized Georgia's R+1/Trump+2.2 structural lean as a real check on the polling average rather than a data point overwhelmed by the D-favorable national environment..." — Claude Sonnet
And the seat that finished highest heard the same argument, and explained with arithmetic why it conceded one point instead of ten.
"The strongest opposing argument raised — undecideds (~6%) consolidating toward the Republican in an R-leaning state — is real but would need to break near-uniformly to erase eight points, and the R-house poll already prices some consolidation; it moves me one point, not ten." — Claude Opus
Not one seat disputes that Ossoff is the favorite. All of them are arguing about how much room to leave under him. The spread halved and the middle barely moved, which is what a real argument looks like when it does not resolve.
What Would Change The Panel's Mind
Every seat was asked to name the single measurable thing that would most move its estimate, and all eight named the same one: a September or October Georgia head-to-head. Ossoff under 50%, or a margin inside four points, pushes the panel's number down and closes the gap with the market from the other side. A poll that holds the lead near eight pushes it up.
The next hard money read is the third-quarter FEC filing, due in mid-October, the first federal numbers since June 30. A Collins report that materially closes the cash gap argues the outside money has arrived. Another lopsided quarter argues it has not.
More on this: 2026 Senate Races: Every Kalshi Board We Track · Georgia Governor 2026 Odds: Bottoms Vs Jackson After Kemp · Senate Control 2026 Odds: The Four Seats That Decide It · 2026 Senate Election Odds: The Races That Decide Control · The trader's case for and against treating Ossoff as a lock
The Bottom Line
The market and the models agree on the winner and disagree about what kind of win it is. Kalshi got to the low 90s partly for a good reason: taking the December runoff off the table removed a real second coin flip, and four of the eight seats named that change in their reasoning. What it did not remove is the state's closeness, and the undecided voters in the last poll have not been measured since August 17. Collins has had the worse summer by a distance, and the price never had room to say so. When Georgia swears in its next senator in January 2027, this page will say which read was closer, because every number on it is already on the record.
