Google to Launch Gemini 3.2 Flash at I/O 2026, Competing with GPT-5.5

According to Beating Monitor, Google plans to release the next-generation lightweight model, Gemini 3.2 Flash, at its I/O conference on May 20. The model’s overall performance is roughly on par with GPT-5.5, though clearly inferior to Anthropic’s Mythos. Abacus.AI CEO Bindu Reddy revealed rumors indicating that Gemini 3.2 Flash achieves 92% of GPT-5.5’s performance on coding and reasoning tasks, while reducing inference costs to just one-fifteenth to one-twentieth of GPT-5.5’s, with most query latencies under 200 milliseconds. She believes Google’s distillation and sparsification techniques are proving highly effective, essentially compressing a state-of-the-art model into a Flash-level version without the typical performance cliff. Preliminary leaks of Gemini 3.2 Flash had already surfaced: in early May, traces of the model were found in iOS app bundles and AI Studio metadata, after which it appeared anonymously in LM Arena evaluations. Early testers reported that the model excels in creative coding tasks and even outperforms Gemini 3.1 Pro on certain benchmarks.