I/O 2026 extended Gemini beyond the chatbot
Google held I/O 2026 on May 19 and presented a strategy focused on bringing AI capabilities to everyday products and building experiences that can carry out tasks, not just answer questions. The main user-facing news was not a single update to the Gemini app, but a collection of models and features spanning Search, YouTube, Workspace, and creative tools.
Two new models took center stage: Gemini Omni Flash, designed to generate content across modalities, and Gemini 3.5 Flash, which Google described as combining advanced capabilities with the ability to take action. The company also discussed agents integrated into Search and the Gemini app. In that sense, a “new chapter for Gemini” describes a product strategy, not a single launch with one shared availability date.
The practical difference is where each capability appears. Some are being integrated into services many people already use, while others are aimed at developers or subscribers and depend on a phased rollout. The announcement does not mean every feature is active for every account: Google gave different timelines and conditions for each one.
Two models with different functions and audiences
Google introduced Gemini Omni Flash as the first model in the Omni family. It said Gemini Omni can generate outputs in different modalities from inputs in any modality; the initial release focused on video generation, with image and text outputs to be enabled later. Omni Flash was announced as available to try that day in the Gemini app, Google Flow, and YouTube Shorts; APIs for developers and enterprise customers were expected in the following weeks.
That does not mean Omni Flash already supported every input-and-output combination Google envisions for the family. The presentation said the rollout would start with video, with other outputs to follow. For app users, the immediate change was the ability to experiment with video creation through the listed products; for technical teams, API access was on a later schedule. Omni’s broader future scope and Omni Flash’s initial availability are different things.
Gemini 3.5 Flash was introduced as the first model in a new family combining advanced capabilities with action, with improvements highlighted in coding, extended tasks, and real-world workflows. In his remarks, Sundar Pichai compared its performance with other models and highlighted its speed. Google also described using it internally alongside an updated version of Antigravity, its agent-oriented development platform.
Features for Search, YouTube, and the Gemini app
In Search, Google announced information agents designed to help with queries that require gathering and organizing data. The idea is to move from a standalone answer to a more structured task within the search engine. Along similar lines, Ask YouTube was presented as a way to explore questions through videos and jump to relevant segments, rather than manually searching through results.
Google said Ask YouTube would begin testing at the time of the announcement, with a broader U.S. rollout planned for mid-2026. That was a forecast announced in May, not proof that the feature had reached every country or account. Access and geographic coverage remain important when interpreting any later news about the tool.
For the Gemini app, Google announced Gemini Spark and Daily Brief as agent experiences. The company included them among the new features intended to bring agents and new experiences into its products.
Workspace: document creation and voice
Docs Live was one of the productivity announcements. Google described it as an experience for creating documents from a spoken explanation and then editing them by voice. The idea is to reduce the need to turn a broad request into a precise written instruction in advance. For people drafting notes, outlines, or work documents, the potential change is to start by speaking and let Gemini structure the content.
The announced availability was not immediate for everyone. Google said Docs Live would roll out to subscribers in mid-2026 and also previewed voice features for Gmail and Keep for the same period. The announcement did not specify that every plan, language, or market would get access at the same time. Subscriber status alone does not determine compatibility; the specific conditions depend on the product and Google’s rollout.
These announcements should be distinguished from earlier Gemini features in Workspace. For example, Docs and Gmail’s assisted writing tool had already expanded to Spanish and Portuguese in May 2024. I/O 2026 introduced a new voice interaction, not the debut of Gemini writing assistance or its original arrival in those apps.
Timeline and scope: what the announcement established
The timeline announced at I/O distinguished features available to try on the announcement date from those planned for later. Gemini Omni Flash was available to try in the Gemini app, Google Flow, and YouTube Shorts; its APIs for developers and enterprise customers were expected in the following weeks. Docs Live, voice features for Gmail and Keep, and a broader Ask YouTube rollout were described with later timelines.
| Announcement | Status or timeline given in May 2026 | Stated scope |
|---|---|---|
| Gemini Omni Flash | Available to try from the announcement | Gemini app, Google Flow, and YouTube Shorts |
| Gemini Omni Flash API | Following weeks | Developers and enterprise customers |
| Ask YouTube | Testing underway; broader rollout planned for mid-2026 | United States |
| Docs Live and voice features for Gmail and Keep | Rollout planned for mid-2026 | Subscribers |
Google announced these timelines in May 2026. To assess whether a feature fits a particular workflow, check the specific product and account.
In practical terms, I/O 2026 pointed to three distinct developments: more multimodal generation, greater integration of agents into Google services, and new ways to create documents with voice. For users, the value depends on actual access and the task at hand; for developers, API availability and integration requirements matter more than demonstrations. The announcements were broad, but they did not make every proposal a universal or simultaneous feature.
How to read the changes alongside earlier announcements
Gemini had already expanded across multiple products before I/O 2026: Google Workspace, Search, and the Gemini app had all received AI features in earlier cycles. May’s news brought models and experiences together with a stronger emphasis on actions, multimodal creation, and voice. That makes the shift an evolution of the platform, not an instant replacement of existing features.
Google presented some announcements as tests, future rollouts, or API access expected in the following weeks. In his remarks, Sundar Pichai described internal use of Gemini 3.5 Flash alongside an updated version of Antigravity and said Omni Flash APIs would reach developers and enterprise customers in the following weeks.
Google introduced Omni Flash for video creation and Gemini 3.5 Flash as a model in a new family focused on advanced capabilities and action. It also announced Docs Live, an experience for creating and editing documents by voice, with a rollout to subscribers planned for mid-2026.