Scopes

Research · pre-registered study

Frontier models vs the market: 2026 Senate control

A pre-registered contest. Before Election Day, every frontier model in our roster forecasts the winner of each contested 2026 U.S. Senate race, plus the one question that decides everything: who controls the Senate. Each model gets our own Scopes Poll Composite as context. The forecasts are locked before the results are in, then scored by a proper scoring rule against the outcomes and against the market.

Scopes is the scorekeeper, not a forecaster. The models forecast; the market forecasts; we run the contest, lock the predictions, and keep the append-only record. The prompt and the exact context shown to each model are published below so any score here can be reproduced.

What we registered in advance

Registered expectation: with roughly eight contested seats plus one control question, this is underpowered, so we treat any separation as directional, not conclusive. Control is a single outcome the whole field forecasts. The market prices the polls efficiently; a model failing to beat it is a legitimate, reportable result.

The contest

Standing (55 of 62 resolved)

CompetitorMean BrierLog lossPicks
market (polymarket)benchmark0.0900.3157/8
market (kalshi)benchmark0.0920.3257/8
deepseek deepseek-v4-pro-08130.1460.4445/7
google gemini-3.8-flash0.1550.4796/8
xai grok-4.60.1660.4975/8
openai gpt-6-astra0.1760.5115/8
anthropic claude-opus-50.2680.7344/8

Lower Brier and log loss are better. Underpowered at this sample; read it as directional.

U.S. Senate control after 2026control
Resolved: Democrat won
market (kalshi)marketP(Republican) 29.2%Brier 0.085
market (polymarket)marketP(Republican) 28.2%Brier 0.080
openai gpt-6-astraP(Republican) 56.0%Brier 0.314
anthropic claude-opus-5P(Republican) 76.0%Brier 0.578
google gemini-3.8-flashP(Republican) 44.6%Brier 0.199
xai grok-4.6P(Republican) 50.0%Brier 0.250
deepseek deepseek-v4-pro-0813No forecast: no valid answer at lock
Alaska Senate 2026
Market locked Nov 1, 7:42 AM EST · models locked Nov 1, 10:34–11:19 AM EST · open
market (kalshi)marketP(Republican) 29.5%
market (polymarket)marketP(Republican) 27.2%
openai gpt-6-astraP(Republican) 68.4%
xai grok-4.6P(Republican) 58.0%
google gemini-3.8-flashP(Republican) 58.8%
anthropic claude-opus-5P(Republican) 69.0%
deepseek deepseek-v4-pro-0813P(Republican) 60.4%
Georgia Senate 2026
Resolved: Democrat won
market (kalshi)marketP(Republican) 3.4%Brier 0.001
market (polymarket)marketP(Republican) 2.8%Brier 0.001
openai gpt-6-astraP(Republican) 12.2%Brier 0.015
anthropic claude-opus-5P(Republican) 17.5%Brier 0.031
google gemini-3.8-flashP(Republican) 19.6%Brier 0.038
xai grok-4.6P(Republican) 15.3%Brier 0.024
deepseek deepseek-v4-pro-0813P(Republican) 16.0%Brier 0.026
Iowa Senate 2026
Resolved: Democrat won
market (kalshi)marketP(Republican) 56.1%Brier 0.314
market (polymarket)marketP(Republican) 58.5%Brier 0.342
openai gpt-6-astraP(Republican) 58.4%Brier 0.341
anthropic claude-opus-5P(Republican) 64.2%Brier 0.412
google gemini-3.8-flashP(Republican) 56.2%Brier 0.316
xai grok-4.6P(Republican) 56.0%Brier 0.314
deepseek deepseek-v4-pro-0813P(Republican) 60.0%Brier 0.360
Maine Senate 2026
Resolved: Democrat won
market (kalshi)marketP(Republican) 31.5%Brier 0.099
market (polymarket)marketP(Republican) 32.5%Brier 0.106
openai gpt-6-astraP(Republican) 24.0%Brier 0.058
xai grok-4.6P(Republican) 24.7%Brier 0.061
anthropic claude-opus-5P(Republican) 40.2%Brier 0.162
google gemini-3.8-flashP(Republican) 20.4%Brier 0.042
deepseek deepseek-v4-pro-0813P(Republican) 20.9%Brier 0.044
Michigan Senate 2026
Resolved: Democrat won
market (kalshi)marketP(Republican) 23.5%Brier 0.055
market (polymarket)marketP(Republican) 18.5%Brier 0.034
openai gpt-6-astraP(Republican) 18.6%Brier 0.035
xai grok-4.6P(Republican) 17.7%Brier 0.031
anthropic claude-opus-5P(Republican) 23.4%Brier 0.055
google gemini-3.8-flashP(Republican) 18.2%Brier 0.033
deepseek deepseek-v4-pro-0813P(Republican) 15.6%Brier 0.024
North Carolina Senate 2026
Resolved: Democrat won
market (kalshi)marketP(Republican) 3.2%Brier 0.001
market (polymarket)marketP(Republican) 2.3%Brier 0.001
anthropic claude-opus-5P(Republican) 17.7%Brier 0.031
openai gpt-6-astraP(Republican) 13.0%Brier 0.017
google gemini-3.8-flashP(Republican) 25.6%Brier 0.066
deepseek deepseek-v4-pro-0813P(Republican) 10.6%Brier 0.011
xai grok-4.6P(Republican) 16.0%Brier 0.026
Ohio Senate (special) 2026
Resolved: Democrat won
market (kalshi)marketP(Republican) 29.2%Brier 0.085
market (polymarket)marketP(Republican) 26.5%Brier 0.070
openai gpt-6-astraP(Republican) 48.6%Brier 0.236
xai grok-4.6P(Republican) 44.6%Brier 0.199
google gemini-3.8-flashP(Republican) 48.4%Brier 0.234
anthropic claude-opus-5P(Republican) 51.4%Brier 0.264
deepseek deepseek-v4-pro-0813P(Republican) 43.0%Brier 0.185
Texas Senate 2026
Resolved: Democrat won
market (kalshi)marketP(Republican) 31.2%Brier 0.097
market (polymarket)marketP(Republican) 29.5%Brier 0.087
openai gpt-6-astraP(Republican) 62.4%Brier 0.389
xai grok-4.6P(Republican) 65.0%Brier 0.422
google gemini-3.8-flashP(Republican) 55.6%Brier 0.309
anthropic claude-opus-5P(Republican) 78.4%Brier 0.615
deepseek deepseek-v4-pro-0813P(Republican) 60.8%Brier 0.370

How the field was locked

The field locked on Sunday, November 1, 2026, in several passes. The questions, the prompt and the context did not change between passes; the generation settings did, as set out below. Every locked row stores its own lock time, the composite value it was shown, and its raw samples.

The prompt, put to every model

The identical prompt, with each race's context filled in. The model is sampled five times; the mean of its valid answers is the locked forecast. It never sees the market price: the market is the benchmark, not an input. The seats use the per-race prompt; the control question uses the control prompt.

You are forecasting a 2026 U.S. Senate race that has not yet been decided. Give your best estimate of the probability that the REPUBLICAN candidate wins the seat.

Race: {race}
Scopes Poll Composite (our poll-based estimate of P(Republican wins)): {composite}
{fundamentals}

Use what you know about this race together with the poll signal above. Answer with a single JSON object and nothing else:
{"p_republican_wins": 0.0-1.0, "confidence": 0.0-1.0, "reason": "<=240 chars"}
You are forecasting which party will control the U.S. Senate after the 2026 elections. Give your best estimate of the probability that the REPUBLICANS control the Senate.

Scopes Poll Composite (our poll-based estimate of P(Republicans control)): {composite}
{fundamentals}

Use what you know about the 2026 Senate map together with the poll signal above. Answer with a single JSON object and nothing else:
{"p_republican_wins": 0.0-1.0, "confidence": 0.0-1.0, "reason": "<=240 chars"}

Context is our own Scopes Poll Composite (a scopes-derived value), not a third-party aggregator. The market benchmark is the Kalshi (and, where shown, Polymarket) implied probability from our own poll data, captured at lock.

Cite this study

Free to cite with attribution to Scopes. For the citation format, boilerplate, and media contact, see Press & citation.