At the May IO Conference, Google unveiled that its premier large model, Gemini 3.5 Pro, initially earmarked for a June release, faced a postponement. During this hiatus, a number of AI specialists at Google departed, taking a toll on team morale. However, the most recent updates reveal that Gemini 3.5 Pro is set to make its debut on July 17th, aligning with the official rollout of DeepSeek V4. The looming clash between these two flagship large models has sparked considerable excitement. The rationale behind the delayed launch of Gemini 3.5 Pro is to harness the power of new pre-training techniques to elevate its performance, rather than relying on mere fine-tuning of the previous model. This approach has yielded remarkable advancements in several key areas, including front-end generation, design aesthetics, UI simplicity, SVG generation, and the refinement of one-time front-end results. Additionally, the model has made strides in game development capabilities. Simultaneously, Google is poised to introduce a new Nnao Banana Pro large model, built upon the foundation of Gemini 3.5 Pro, to go head-to-head with GPT-Image 2.