Google has pushed its AI strategy further into everyday products with two new models that target very different uses: Gemini 3.5 Flash for general intelligence and Gemini Omni for video generation. The move shows that Google is not limiting its latest AI work to chat interfaces, but is instead weaving it into Search, consumer apps, and creator tools.
Gemini 3.5 Flash is now the first model from the 3.5 family to be released by Google. The company says it combines flagship-level intelligence with the speed expected from the Flash line, and it has been positioned as the default model for Gemini. That makes it the core engine for daily AI use across Google’s consumer-facing services.
Google also describes Gemini 3.5 Flash as its strongest model for agentic and coding tasks. In the company’s claims, it outperforms Gemini 3.1 Pro on demanding coding and agentic benchmarks, while also showing stronger multimodal understanding.
That multimodal capability matters because modern AI systems are increasingly expected to work across text, images, audio, and other inputs in a single workflow. Google’s decision to place Gemini 3.5 Flash inside the Gemini app and AI Mode in Google Search suggests that the model will influence both direct AI use and search-assisted experiences.
A broader push into video creation
On the media side, Google introduced Gemini Omni as a new model focused on turning varied inputs into video. The model can use images, audio, video, and text together as prompts, then generate high-quality video that Google says is grounded in Gemini’s real-world knowledge.
The design is not limited to simple text-to-video generation. Google says the model is built to process richer source material and shape it into more complex video output, which places multimodal understanding at the center of the creative process.
Once a video is generated, it can be edited through conversation. That approach allows users to refine results without restarting from scratch every time they want a change.
The first model in this family is Gemini Omni Flash. Google says it lets users adjust specific elements in a video, or even alter the entire video, and then keep refining it across multiple rounds while preserving the context of the original scene.
Google also says the model has improved intuitive understanding of style, kinetic energy, and fluid dynamics. Those capabilities are aimed at making motion and physical behavior look more convincing, especially in scenes that are traditionally difficult for AI video systems to render well.
Rolling out across Google’s ecosystem
Gemini Omni Flash is available globally for subscribers to Google AI Plus, Pro, and Ultra plans. Access is provided through the Gemini app and Google Flow, while Google has also started rolling it out for free to YouTube Shorts and YouTube Create users.
The wider rollout signals that Google wants the video technology to reach creators beyond premium AI subscribers. The company is also allowing users to use their own voice and Avatars, which are digital versions of the user, adding a more personal layer to AI-generated content.
To help identify synthetic media, every video made with Omni will include a SynthID digital watermark. That label is becoming more important as major platforms expand support for AI-generated content and creators use these tools at greater scale.
Taken together, Gemini 3.5 Flash and Gemini Omni show two parallel directions in Google’s AI development. One strengthens the company’s core model for search, coding, and agentic tasks, while the other pushes generative AI further into flexible, editable video creation for everyday users and creators.
Source: www.gsmarena.com
