Google’s Google I/O conference opening today drives the near-certain trader consensus around a May 19 release for Gemini 3.2, the latest iteration in the company’s large language model lineup. Recent leaks of model identifiers in AI Studio logs, Gemini app metadata, and LM Arena benchmarks mirror the exact pattern that preceded the February 19 launch of Gemini 3.1 Pro, while insiders indicate a Flash variant optimized for speed and multimodal tasks will debut without a new Pro model. This fits Google’s stepped-up cadence of smaller, frequent updates aimed at narrowing gaps with OpenAI and Anthropic systems. A last-minute delay from integration issues or a shift to a May 20 announcement at the event remain the only realistic ways the market could move.