Google has launched two new AI models that fundamentally change how developers create and iterate on video content by combining ultra-fast image generation with conversational video editing in a single workflow. On June 30, 2026, Google released Nano Banana 2 Lite for rapid image generation and Gemini Omni Flash for conversational video editing, making high-volume creative production commercially viable for the first time through an integrated pipeline that previous AI tools never offered.

What Makes This Different From Other AI Video Tools?

Every major AI video tool released before Gemini Omni Flash operated in a generate-and-export paradigm: a user submits a prompt, the model renders a clip, and if changes are needed, the user re-prompts from scratch or switches to a separate editing application. That workflow makes AI-generated video expensive to iterate on in practice, regardless of per-second pricing.

Gemini Omni Flash breaks that pattern through a combination of architecture and API design. The model is built on Gemini's multimodal reasoning engine, which reasons across text, image, audio, and video inputs simultaneously rather than stitching together separate pipelines. The practical result is the Interactions API, which maintains session history across sequential edits. A developer can generate a 10-second video clip from an image reference, ask the model to adjust the lighting and re-render, then ask it to swap a background element, all within one session with the model retaining context from each prior turn.

"The next step towards the progression of combining the intelligence of Gemini with the rendering capabilities of our media models," explained Nicole Brichtova, product management director at Google DeepMind.

Nicole Brichtova, Product Management Director, Google DeepMind

How to Build an End-to-End Creative Workflow With These Tools

• Image Generation: Use Nano Banana 2 Lite to generate a text-to-image output in four seconds at $0.034 per image, making it fast enough to embed inside a live design tool or e-commerce configurator where users are waiting for results.
• Video Animation: Pass the generated image directly to Gemini Omni Flash to animate it, then refine the result through plain-language commands without re-prompting from scratch or switching applications.
• Iterative Editing: Continue refining video through up to three sequential edits within a single session using Google's Interactions API, adjusting camera angles, swapping characters, and relighting scenes through conversational commands.

The four-second latency for image generation is what changes the category calculus. Prior image generators operated on timescales that put them outside interactive loops, forcing developers to wait, batch results, and adjust. At four seconds, image generation becomes fast enough to embed inside a live creative workflow.

"When generation is faster than ideation, creators stay inside the work rather than breaking flow to wait on a progress bar," said Logan Kilpatrick, who leads Google AI Studio and the Gemini API.

Logan Kilpatrick, Lead, Google AI Studio and Gemini API

What Are the Specific Capabilities and Pricing?

Nano Banana 2 Lite is the fastest and lowest-cost model in Google's four-tier Nano Banana image family. Despite its speed focus, Google states that the model maintains reliable prompt adherence, consistent character rendering across multiple generations, and legible text inside images, the three capabilities most critical to advertising and marketing use cases. The model sits at number five on the public Arena image-generation leaderboard, behind OpenAI's gpt-image-2 and Microsoft's MAI-Image-2.5.

Gemini Omni Flash is priced at $0.10 per second of video output, which matches Google's Veo 3.1 Fast pricing. However, Google explicitly distinguishes the two products: Veo 3.1 excels at high-quality one-shot clip generation, while Gemini Omni Flash is designed for iterative, conversational workflows that combine multiple asset types. Currently, each output can be up to 10 seconds long, with longer durations coming soon.

Both models incorporate SynthID digital watermarking to ensure content transparency and regulatory compliance. The models are available via Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform, with gradual rollout across consumer-facing products such as AI Mode in Search, the Gemini app, NotebookLM, and Google Photos.

What Are the Current Limitations?

Google is transparent about Gemini Omni Flash's current limitations in its launch documentation. Audio reference uploads are not yet supported in the Gemini API. Video references of up to three seconds in duration are accepted by the API schema but are not correctly processed by the model at this time. Character consistency across scene changes and panning movements has documented gaps. Google recommends treating the current release as a prototyping tool for developers rather than a production-ready service.

The model also declines to generate or edit video involving real people's names or likenesses. When such a request is submitted, the model returns an input-blocked message. The filter is consistent with Google's Responsible AI principles and limits deepfake risk, though it also rules out certain legitimate creative applications such as historical reconstructions involving named individuals.

Who Is Already Adopting These Tools?

Enterprise adoption is already underway. WPP has integrated Gemini Omni Flash into its WPP Open agentic platform to provide more controlled AI content production at scale for clients, with teams testing asset localization, product swaps, and dynamic style transfers. Adobe has announced plans to bring both Nano Banana 2 Lite and Gemini Omni Flash into Adobe Firefly.

"Build on Adobe's strategy to deliver our pro-grade tools and the industry's top creative AI models in a connected workflow, giving creators flexibility and control over how they bring their creative ideas to life," stated Matt Chotin, senior director of product at Adobe.

Matt Chotin, Senior Director of Product, Adobe

The dual launch matters most as a unified pipeline. Developers can pass an image generated by Nano Banana 2 Lite directly to Gemini Omni Flash to animate it, and can then continue refining the result through plain-language commands for up to three sequential edits within a single session. That chain is what no prior AI media stack offered at this price: a high-speed image generator and a stateful, conversational video editor unified in one workflow.