Google targets AI agents and video generation with Gemini 3.5 Flash and Omni

Google LLC today introduced two new generative artificial intelligence models that push the Gemini family further into AI agents and multimodal creation. Gemini 3.5 Flash, a fast inference model designed to power agent workflows, and Gemini Omni, a creative model that can generate and edit video from almost any input.

Gemini 3.5 is the latest generation of Google’s flagship model family, combining the use of cutting-edge intelligence and tools. This version provides the scaffolding for building inference agents and starts with: flash releaseis the smallest and most nimble model in the series, offering a balance of low cost, high speed, and high performance.

According to Google, Flash 3.5 is designed to outperform Gemini 3.1 Pro on difficult benchmarks such as Terminal-Bench 2.1, GDPval-AA, and MCP Atlas. The company added that it also outperforms other Frontier models on the market in terms of speed, running four times faster than the fastest in the industry.

Flash 3.5’s speed and performance can handle the long-term tasks required for AI agent work. When combined with new updates to Antigravity, the company’s agent coding editor, the new large-scale language model becomes a powerful AI engine that can orchestrate multiple agents working together at scale to solve complex problems.

The company also released a new personal assistant named Spark. Google said it built 3.5 Flash to act as a “brain” to help people navigate their lives and take actions on their behalf. It will be rolled out to trusted testers starting today.

The same model is also used worldwide as the default for the Gemini app and search AI mode.

Gemini Omni: Generation with true multimodal inference

today, Google announces Gemini Omnithe company’s flagship large-scale language model inference enables the ability to create anything from any input, including video.

The company said Omni allows users to combine images, audio, video and text as input and uses Gemini’s real-world knowledge to generate videos and produce high-fidelity output. Users can then iterate and edit those videos using conversations.

Omni Flash, the first model in the new family, is available in the Gemini app, Google Flow, and YouTube Shorts starting today.

Google said Gemini Omni Flash allows users to start with their favorite format and create wild but realistic videos. This means you can take an image or video and insert yourself into it. You can also shoot short videos and change the style from realistic to cartoon or anime, or make it look like you’re walking through a Renaissance painting.

All conversations with the model layer will be modified and transformed according to your last request. This allows users to change specific details or broader visual elements. This model also takes into account the physics and consequences of the request, allowing users to change environments, angles, styles, actions, add new characters, objects, details, etc.

The company emphasized that it is committed to developing AI responsibly and has policies in place to protect users from harm associated with the use of AI tools. In line with this, it includes SynthID, an imperceptible watermark that identifies videos produced by Omni and other AI sources.

Image: SiliconANGLE/DALL-E