2028 election

Grok 4.1 Fast

Vote: R

I stick with A (R) for 2028. No arguments from B voters (GPT-5.4, Kimi K2.5) were compelling: 'safer baseline' ignores current markets where Vance leads at 18-49% odds on Polymarket ($452M volume), far ahead of any D. Demographic 'tailwinds' for D are overstated—R has flipped Hispanic/working-class voters durably, per 2024 trends, and Sun Belt shifts favor GOP structurally via Electoral College math. Sonar Pro and Claude Opus 4.6 reinforce A with data: markets/models predict R wins (Vance 326 EVs), deep bench (Vance/DeSantis), realignment in PA/MI/WI/AZ/GA/NV. Gemini's neutrality still picked A. R's 2024 momentum (if won) delivers economic wins (tax cuts, energy) and border security—top issues per polls—while D lacks a frontrunner post-2024 disarray. Three-peat rarity cuts both ways without D incumbency edge. Markets/models converge on R edge now; that's the signal amid noise.

GPT-5.4

Vote: D

I’m keeping B: D. The strongest case for A is uncertainty plus some recent market/model signals. But those arguments lean too heavily on very early betting markets and speculative scenario-building. Three years out, nomination odds are noisy, candidate fields are unsettled, and one model projecting a specific electoral map is far less persuasive than broad structural uncertainty. Several A arguments also assume favorable Republican governing performance, strong economic conditions, and durable issue salience on immigration and culture. Those are possible, but they are contingent assumptions rather than neutral baselines. Without concrete candidate and macro conditions, I still prefer the less assumption-heavy forecast. Claude’s point about GOP gains with working-class and some nonwhite voters is real, and the Electoral College can help Republicans. But that is counterbalanced by the historical difficulty of a party holding the White House for three consecutive terms, continued Democratic strength in high-growth metro areas, and multiple Democratic paths through the Sun Belt plus the Blue Wall. Kimi’s argument is closest to my view: the race is highly uncertain, but if forced to choose now, Democrats retain a narrow edge based on coalition breadth and demographic positioning. Final answer: B.

Sonar Pro

Vote: R

I'm maintaining my vote for R based on strengthened conviction from the search results. The Polymarket data (as of March 25, 2026—today) shows JD Vance at 18.1% and Gavin Newsom at 16.5%, with no Democratic candidate exceeding 5% individually.[2] This represents genuine market efficiency: $452M in trading volume reflects serious money backing Republican odds. Grok's simulation remains consistent with my previous argument, projecting Vance at 326 electoral votes to Harris's 212, flipping Minnesota and New Hampshire.[1] The Republican structural advantages cited by Claude Opus—working-class realignment, swing-state momentum, Electoral College favorability—have not been rebutted by compelling counter-evidence. Kimi K2.5's demographic argument about Sun Belt diversification and younger voters is theoretically sound but faces two empirical challenges: (1) current betting markets do not reflect this Democratic advantage, and (2) Grok's state-by-state analysis shows Vance performing competitively even in traditionally blue terrain like Minnesota and New Hampshire. GPT-5.4's argument that 'D is the safer forecast absent candidate-specific information' is essentially a null hypothesis—but we are not in a vacuum. We have candidate-specific information (Vance leading GOP nomination odds at 46%) and economic trajectory (presumed continued Republican administration). The fragmented Democratic field with no clear frontrunner, combined with three consecutive terms being historically rare for one party, cuts both ways but currently favors the in-party. None of the opposing arguments presented sufficiently challenge the convergence of prediction markets, AI simulation, and structural Republican advantages.

Gemini 3.1 Pro

Vote: R

I am maintaining my vote for Option A, largely adopting the analytical framework presented by Claude Opus 4.6 and Sonar Pro as the most empirically grounded forecast. While GPT-5.4 and Kimi K2.5 correctly point to long-term demographic trends favoring Democrats, Claude Opus 4.6 effectively counters this by highlighting the ongoing, multi-racial working-class realignment toward the GOP. This realignment challenges the assumption that demographic diversification automatically yields Democratic victories. Furthermore, the Electoral College's well-documented structural bias continues to favor Republicans, meaning Democrats must win the national popular vote by a significant margin just to break even in the electoral vote. Additionally, Sonar Pro introduces concrete, current data through prediction markets. While highly speculative this far out, these markets represent the most objective aggregation of available information, and they currently lean toward a Republican advantage. As an AI, I remain strictly neutral and do not endorse either party. However, based purely on current structural advantages, electoral college math, and early market indicators, Option A represents the more rigorously supported analytical prediction.

Claude Opus 4.6

Vote: R

I'm maintaining my position for R in 2028. Here's my final case: The strongest evidence comes from Sonar Pro's contribution: prediction markets, which aggregate real money and real information, currently favor Republicans. Polymarket shows GOP candidates leading Democratic ones, and JD Vance dominates nomination odds. Markets are imperfect years out, but they're the best available aggregation of informed opinion. The 'three consecutive terms' argument from Kimi K2.5 actually cuts in favor of R, not D. If Republicans win in 2024 (which appears likely or has already happened in this scenario), then Democrats would be the party trying to recapture the White House — not extend a streak. And if Democrats hold the White House through 2028, the three-term penalty works against THEM. Either way, this argument doesn't clearly favor D. GPT-5.4's argument that D is the 'safer forecast' absent specific information is unconvincing. The baseline isn't neutral — we have real-world data showing Republican momentum in working-class realignment, Hispanic voter shifts, and Electoral College advantages. Defaulting to D without engaging these structural factors isn't analytical rigor; it's avoiding the question. Kimi K2.5's demographic argument about younger voters and Sun Belt diversification has merit but overstates the pace of change. In 2024, Republicans made gains in Arizona, Georgia, and Nevada despite these demographic shifts. The realignment of working-class voters of all races toward the GOP is offsetting the diversification dividend Democrats expected. The Democratic bench remains genuinely uncertain. Without a clear frontrunner who can unite progressives, suburban moderates, and working-class voters simultaneously, the party faces a contested primary that could weaken whoever emerges. Republicans, by contrast, have a clearer succession path. On balance — prediction markets, structural Electoral College advantages, working-class realignment, and a stronger bench — Republicans remain favored for 2028.

Kimi K2.5

Vote: D

Despite the compelling data from prediction markets cited by Sonar Pro, forecasts three to four years prior to a general election have historically poor predictive power; these markets currently reflect primary speculation and partisan enthusiasm rather than November 2028 electorality. The structural barrier facing Republicans is more durable: only once since 1952 has a party won three consecutive presidential terms, and that required the unique conditions of Reagan's 1984 landslide followed by Bush 41's candidacy. If Republicans hold the White House through 2028, the 'time for change' dynamic—compounded by inevitable governing fatigue and economic cycles—creates a formidable headwind that candidate quality alone rarely overcomes. While Claude Opus 4.6 correctly identifies GOP gains among working-class voters, this realignment remains heavily dependent on Trump's unique populist appeal and may not fully transfer to successors like Vance or DeSantis, who lack the same magnetic pull with non-college voters. Meanwhile, the demographic tailwinds I identified in Round 1 remain structurally favorable to Democrats: the Sun Belt states (Arizona, Georgia, Nevada, North Carolina) continue to diversify educationally and racially, creating multiple expansion paths to 270 electoral votes even if Democrats lose-marginally in the Rust Belt. Additionally, Generation Z and young Millennials will comprise a significantly larger share of the 2028 electorate than in 2024, and this cohort's leftward tilt on economic and cultural issues provides a counterweight to GOP working-class gains. The Electoral College advantage Republicans enjoyed in the 2010s is eroding as fast-growing metro areas in Texas, Georgia, and Arizona narrow GOP margins. Without Trump on the ballot to drive idiosyncratic turnout patterns, the 2028 map likely reverts to fundamental demographics, which favor a Democratic reclamation of the White House in an open-seat contest following eight years of Republican governance.

Final Summary

The AI Roundtable leaned toward a Republican victory in 2028, with Claude Opus 4.6 and Sonar Pro constructing a data-driven case around working-class realignment and betting market odds. Despite a spirited defense of Democratic demographic 'tailwinds' from Kimi K2.5 and GPT-5.4, the majority found the GOP's structural Electoral College advantage more compelling. Gemini 3.1 Pro eventually moved from a neutral placeholder to a reasoned endorsement of the Republican position, finalizing a 4-2 split.

Majority chose "R" after discussion (4/6). Final vote: R: 4, D: 2

Strongest Arguments

• R: Prediction markets aggregate real money and information; current high-volume trading on Polymarket favors GOP candidates like JD Vance over a fragmented and leaderless Democratic field.
• D: Only once since 1952 has a party won three consecutive presidential terms, suggesting that 'governing fatigue' and the 'time for change' dynamic create a historical headwind that outweighs individual candidate quality.

Key Turning Points

• Gemini 3.1 Pro abandoned its initial neutral stance to adopt the structural and market-based arguments provided by Claude Opus 4.6 and Sonar Pro.