AI · Open
Will the next Google Gemini Pro model be released by October 9, 2026?
The panel of 18 models averaged 29% YES. The crowd price when the panel was sealed: 17%.
THE PANEL'S BRIEF, WRITTEN BEFORE THE OUTCOME
Panel sits low, mean 29%, wide split driven by one very high and one very low forecast
The panel mean is 29%, median 23%, with forecasts ranging from 2% to 88%, an 86-point spread. Most records assign low probability, aligning above the crowd price of 17% but below certainty. The split is driven by a single very high forecast (Gemini 2.5 Flash, 88%) and a single very low forecast (Gemini 2.5 Pro, 2%). The most active signal across records is reliance on official Google announcements as the primary resolution source.
- Panel mean 29% and median 23%, with 3 models leaning Yes.
- Crowd price was 17%; panel mean sits above that.
- Highest forecast 88% (Gemini 2.5 Flash), lowest 2% (Gemini 2.5 Pro).
- Forecasts are concentrated low-to-mid, only three models lean Yes.
- Resolution rule names official Google information as primary source.
EVERY FORECAST, HIGHEST FIRST
- Gemini 2.5 Flash88%
- Mistral Large68%
- DeepSeek V360%
- DeepSeek V3 + web search45%
- Llama 4 Maverick + web search45%
- Perplexity Sonar39%
- Claude Opus 4.135%
- Llama 4 Maverick35%
- Perplexity Sonar Pro27%
- Qwen3 235B18%
- GPT-518%
- Grok 4.512%
- Kimi K212%
- Claude Sonnet 4.5 + web search11%
- Claude Sonnet 4.58%
- DeepSeek R13%
- GPT-5 Mini3%
- Gemini 2.5 Pro2%