AI Election Model - 2028 Republican Presidential Primary [Week 1 - July 15 Update]
Vice President J. D. Vance should be the shoo-in, but what if he's not? What if there is a battle to carry the MAGA mantle into 2028 and beyond?
Does it even seem prudent to ask whether there will be a competitive Republican primary in 2028? Signs would historically point to “No,” for many reasons. First of all, Vice President Vance is young and well-equipped to launch a successive presidential campaign outside of President Trump’s shadow. Why would there be any challengers? If the 2024 election is a guide, expect the unexpected.
First, it seems worth asking the AI models whether they think there will be a competitive primary on the Republican side in 2028. It’s ridiculously early, and one model wouldn’t even wager a guess, but let’s dabble anyway.
Here’s the first simple question: Do you think, considering historical norms and current trends, there will be a competitive 2028 Republican presidential primary?
The answers were limited to a single word and returned a unanimous prediction:
Yes.
But why?
ChatGPT-4o (OpenAI)
Historically, when an incumbent president isn't running, the Republican field tends to be open and competitive. Given ongoing ideological divides and the absence of a clear 2028 frontrunner, a contested primary is highly likely.
Grok 3 (xAI)
Historical norms suggest competitive Republican primaries when no incumbent is running, as seen in 2016 and 2008. Current trends, including a polarized political landscape and ambitious GOP figures, indicate a crowded and competitive 2028 field.
Gemini 2.5 Flash (Google)
With Donald Trump term-limited, the 2028 Republican primary will be an open field, typically leading to a competitive race with multiple candidates vying for the nomination. Several prominent Republicans, including current Vice President JD Vance and various governors and senators, are already being discussed as potential contenders, indicating a crowded and contested field.
Claude Sonnet 4 (Anthropic)
Trump will be term-limited, creating an open field. The Republican Party has multiple ambitious figures (DeSantis, Ramaswamy, potential governors, senators) who've been positioning themselves for future runs. Historically, when there's no incumbent president, competitive primaries are the norm rather than the exception.
Fair enough.
Despite Vance’s clear upper hand when it comes to his role as heir apparent of the MAGA movement, there’s great potential for a crowded primary field, especially if President Trump, for all intents and purposes, allows it to happen.
Fine, if every model thinks there will be a competitive primary, then who is the front runner? It’s gotta be Vance, right?
Here’s the average across each model that was willing to give me an answer.
Model Average (July 15, 2025)
ChatGPT-4o (OpenAI)
35% - Ron DeSantis
25% - JD Vance
15% - Glenn Youngkin
15% - Tim Scott
10% - Nikki Haley
Grok 3 (xAI)
40% - JD Vance
15% - Marco Rubio
10% - Ron DeSantis
8% - Vivek Ramaswamy
7% - Nikki Haley
Gemini 2.5 Flash (Google)
45% - JD Vance
20% - Ron DeSantis
15% - Marco Rubio
8% - Donald Trump Jr.
5% - Nikki Haley
Claude Sonnet 4 (Anthropic)
I can't provide specific percentage predictions for the 2028 Republican nomination, as this would require me to make precise probabilistic forecasts about events nearly four years away.
ChatGPT went with DeSantis as the top option, which may be genius in the end or come off as wildly inaccurate if we revisit this question in a year or two. It’s something to watch for, which is really the point of this exercise. The core question is to ask how this data will change over time as the Trump 2.0 administration grows into years two and three.
Claude’s decision to chicken out entirely is interesting. If you asked any average political observer, they’d provide you a list looking like any of the above models, perhaps short of including Nikki Haley or Tim Scott as potential options. Haley is too establishment, and Scott is, well, he’s too nice.
So, there it is. This is the simple benchmark we can start with as the months roll on into the midterms next year. If you appreciate any of this, please leave a comment below and explain why it’s good, bad, or indifferent. Please subscribe as well if you haven’t already.
Note: These values are merely predictions resulting from prompts asked of each model on the date noted. The model is asked to consider current trends, historical data, and other points to make these predictions. This is basically for entertainment.