Rumors about the Gemini Pro model are swirling again, with smart money in the AI sector predicting the new Pro model will be released before November.

According to PolyBeats monitoring, 0x7365 placed $1.1k on "Yes" for the question "Will the next-generation Gemini Pro model be released before November?" with an average purchase probability of 71.6%, and the current "Yes" probability stands at 69.5%.

0x7365 invested $1.1k, with the most relevant sector in this market being AI, where the sector net profit is $41.0k. Across 639 settled trades in this sector, the win rate is 355/639 (56%), including 132 trades with buy prices below $0.8 and sell prices above $0.95. Within a similar cost price range ($0.651-$0.8), the median historical investment amount is $565.

On September 5, X blogger Lyra (lyraxana) posted that Google, after a long gap, released the first internal checkpoint of the upcoming Pro model. He expects the Pro to be publicly released in October, along with a new Flash-Lite and an updated NB2Lite in September. The account's last similar prediction was on August 31, stating that gemini-3.8-flash had been deployed and would be released the next day, with the actual public release on September 2, a one-day discrepancy.

Netizens speculate the product could be Gemini 4 Pro, while Gemini 3.5 Pro may have been shelved. Google has updated training data to improve model performance in code generation and software engineering tasks, but test results still fall short of internal targets. Some testers have already accessed the model through Google's relevant platforms, and the team continues to evaluate its capabilities in long-horizon tasks, agent execution, and multi-step reasoning.