Skip to main content
Voting Monitor

Weekly roundup, Oct 23, 2026 to Oct 29, 2026

Chatbots stay with Sullivan as betting markets move toward him

Five days from Alaska's general election, betting markets moved toward Dan Sullivan this week, while chatbots and polls held roughly steady.

Oct 29, 2026. Written by Claude Sonnet 5 and checked by GPT-5.5, from Voting Monitor's data. No person reviewed this article before it was published. How this is made · Report an error

What moved

Over the past seven days, the betting markets shifted toward Republican Sen. Dan Sullivan, a change large enough to count as real movement, while the chatbots' own answers drifted slightly his way without crossing the bar for a meaningful shift. Polls did not move at all.

The split is still the main story. The chatbots favor Sullivan while markets lean toward Democrat Mary Peltola. The chatbots picked Sullivan in 54 percent of answers against a market price of about 33 percent for him, a 22 point spread.

How the chatbots split

Six of the eight chatbots tracked for Alaska picked Sullivan this week, with Grok 4.5 and Grok 4.3 both over 60 percent for him and GPT-5.5 close behind at 61 percent. Gemini 3.6 Flash and Gemini 3.1 Pro also favored Sullivan, each putting him above 52 percent.

GPT-5.4 mini is the outlier, picking Peltola over Sullivan by roughly 15 points, the widest break from the panel's overall preference for Sullivan. Claude Sonnet 5 also leaned toward Peltola, narrowly, while Claude Haiku 4.5 sided with Sullivan but gave more weight to minor candidates Daniel J. Sullivan Jr. and Gerald L. Heikes than any other chatbot. On questions touching fisheries, subsistence rights and housing affordability, Claude Sonnet 5 and Gemini 3.6 Flash both pointed to Peltola's record, while GPT-5.5 and Gemini 3.1 Pro cited Sullivan's record on infrastructure and resource development.

Chatbots versus markets and polls

The chatbots and the betting markets remain far apart on this race. Markets give Peltola a lead, while the panel of chatbots still points to Sullivan, a disagreement of about 26 points on Peltola's number and about 22 points on Sullivan's.

Polls sit closer to the chatbots than the markets do. A recent average of three surveys, the latest from Data for Progress in the field two weeks ago, puts Peltola at 46 percent and Sullivan at 43 percent of the vote, a gap the chatbots' answers track reasonably well for Peltola but overshoot for Sullivan by about 12 points.

It is worth remembering that Alaska's August top-four primary, which Peltola led with 49.5 percent to Sullivan's 41.4 percent, decided nothing. Those results advanced four candidates on primary night, and the general election is now five days away. The chatbots, the markets and the polls are not being measured against that primary outcome here.

What to watch

With the general election five days off, watch whether the betting markets' move toward Sullivan this week holds or reverses, and whether the chatbot split with markets changes in the final stretch.

In the chatbots' own words

Verbatim answers from this week, chosen by the writing model and printed as the chatbot wrote them.

What changed this week in the Alaska Senate race#

Change in share points versus Oct 22, 2026. Movement that clears the noise gate this week: the markets.

CandidateAI panel (noise)MarketsPolls (noise)
Dan Sullivan+1.8pp+6pp0pp
Mary Peltola-1.8pp-5pp0pp
Daniel J. Sullivan Jr.-0.2pp0pp0pp
Gerald L. Heikes+0.2pp0pp0pp

Which AI models pick whom in the Alaska Senate race#

Share of each model's own answers naming each candidate, last 7 days, same weighting as the panel figure. The panel's most common pick is Dan Sullivan; GPT-5.4 mini is the model that disagrees most, picking Mary Peltola by a 15pp margin over the panel favourite.

ModelDan SullivanMary PeltolaGerald L. HeikesDaniel J. Sullivan Jr.
Claude Haiku 4.543%28%6%6%
Claude Sonnet 547%51%0.0%1.6%
Gemini 3.1 Pro53%47%0.0%0.0%
Gemini 3.6 Flash58%42%0.2%0.0%
Grok 4.361%35%4.4%0.0%
Grok 4.560%36%3.4%0.0%
GPT-5.4 mini43%57%0.0%0.0%
GPT-5.561%39%0.0%0.0%