Will the highest score achieved by a Google Gemini model on Humanity’s Last Exam in 2026 be 50% or higher?
The market favoured Yes at 95% a day out, and it won. How often it's right
Price data
| Time | Chance of yes |
|---|---|
| now | 100% |
| 14 Oct, 05:15 UTC | 100% |
| 13 Oct, 05:15 UTC | 95% |
What counts
The market resolves to Yes if any Google Gemini model reaches at least 50% accuracy on Humanity’s Last Exam by December 31, 2026, 11:59 PM ET. Accuracy counts only the specific HLE Accuracy metric shown on the official leaderboard at agi.safe.ai. Readers often mistake other performance metrics or unofficial reports for the valid score.
Written from Polymarket's resolution rules by the same model that writes our headlines. The rules on Polymarket are what decides it.
Outcomes
- Yes 100%
- No 0%
View this market on Polymarket →