Google DeepMind

Google DeepMind's release of Gemini 3.5 marks their latest major model release, introducing a new series of models that combine frontier intelligence with action. Designed to help developers efficiently build more capable, intelligent agents, the Gemini 3.5 series handles everything from completing everyday tasks to solving the most complex challenges, ensuring users can easily discover the exact right model for their needs.

As a major milestone, this release required a distinct visual system to clearly set it apart while staying true to the existing Gemini brand. Working alongside the Google DeepMind team, we supported them in finding a scalable way to apply a cohesive brand identity across an ever-expanding, diverse range of models—from Flash to Audio.

To streamline this process, we created a custom generative tool that empowers the team to select any icon from their vast library, apply a signature glowing particle effect, and export the final asset directly. Each icon can be exported with a variety of motion presets, as well as in both day and night modes. Using this platform makes the ongoing release of new models far more efficient and unifies how the brand appears across all touchpoints.