Google Ships Gemini 3.5 and Android XR Glasses at I/O 2026

*Google used its annual developer event to push Gemini models deeper into search, video tools, and new extended-reality hardware.*

Gemini Updates

Google introduced Gemini 3.5 alongside a new model called Gemini Omni. Omni accepts any input type and generates output across formats, with improved handling of physical simulations such as gravity and motion. It merges core Gemini reasoning with separate generative systems, including Nano Banana and Veo. The model supports video editing through natural language: users upload a clip and describe changes to specific elements, and the system applies those edits while preserving the rest of the scene. Initial release focuses on video, with further modalities planned.

Search and Agent Changes

Search now incorporates more agent-style behaviors that complete multi-step tasks without additional prompts. The updates aim to reduce the number of follow-up queries users must issue. Google also described plans to embed similar agents across other products, though rollout timing varies by service.

Hardware and Platforms

New Android XR glasses were shown, positioned as a consumer device arriving this fall. The glasses combine camera input with Gemini processing for real-time assistance. Developers received early access tools tied to the same Gemini stack that powers the glasses. No pricing or exact availability dates were released during the event.

Reactions

Coverage from multiple outlets noted the breadth of announcements but offered limited third-party testing data at the time of the keynote. Direct comparisons to competing AI products were not detailed in the presentations.

Why it matters

The moves tighten Google’s grip on the AI layer that sits between users and its core services. Developers who build on Gemini gain access to stronger video and agent capabilities, while consumers will see incremental changes in search and new hardware options later this year. The approach keeps pressure on rivals to match both model quality and distribution reach.

---

Sources:

• 9to5Mac
• MacRumors
• Wired

{
  "excerpt": "Google detailed Gemini 3.5, a new Omni model for video editing, AI agents in search, and Android XR glasses at its 2026 developer conference.",
  "suggestedSection": "ai",
  "suggestedTags": ["google-io", "gemini-ai"],
  "imagePrompt": "An abstract workspace with floating translucent panels showing fragmented video frames and geometric motion trails suspended above a dark reflective surface. Soft directional light catches the edges of the panels and trails. muted color palette, cinematic lighting, 16:9"
}