Google's Foundation Models: Gemini, Gemma, Imagen, and Veo Explained

Google's foundation models are large models trained on broad datasets that provide general-purpose generative capability, and the four the exam names are Gemini, Gemma, Imagen, and Veo. A foundation model is a starting point: rather than building a model for one narrow task, a business takes a foundation model and applies it, through prompting, grounding in company data, or tuning, to many tasks.

The four models divide the landscape along two lines. The first is modality, the type of content a model understands or produces: Gemini works across text, images, audio, video, and code; Imagen produces images; Veo produces video. The second is openness: Gemini, Imagen, and Veo are managed models accessed as a service, while Gemma is an open model whose weights can be downloaded and run wherever the organization chooses.

All four are available to builders through Model Garden on Agent Platform, and Gemini also powers Google's end-user applications such as the Gemini app and Gemini for Google Workspace.

Gemini is Google's flagship foundation model family, and its defining strength is that it is natively multimodal: it was built from the start to reason across text, images, audio, video, and code together, rather than handling one type of content at a time. Paired with a long context window, this means Gemini can take in large, mixed inputs, a lengthy contract, a set of product photos, a recorded meeting, and reason about them in a single request.

Its strengths translate directly into business use cases. Gemini can summarize and answer questions over long documents, power customer-facing assistants that hold natural conversations, analyze charts and screenshots, draft and review code for development teams, and act as the reasoning engine inside agents that complete multi-step tasks. When an exam scenario involves understanding, conversation, analysis, or any mix of input types, Gemini is the answer.

A concrete example: an insurance company processing claims receives photos of vehicle damage, a written incident description, and a phone-call recording. Because Gemini is multimodal, one model can review all three, extract the relevant facts, flag inconsistencies between the photos and the description, and draft a structured claim summary for the adjuster. Before multimodal foundation models, that workflow would have required separate systems for images, text, and audio, plus custom work to combine their outputs.

Gemma is Google's family of lightweight open models, built from the same research and technology behind Gemini. Open means the model weights are published: an organization can download Gemma, run it on its own infrastructure, on a chosen cloud environment, or even on capable local devices, and fine-tune it on its own data. That stands in contrast to Gemini, which is accessed as a managed service.

Gemma's strengths follow from that openness. It gives organizations control, over where the model runs, how data flows, and how the model is customized, and cost efficiency, because lightweight models need far less compute to run than flagship models. Developers can inspect, adapt, and specialize it deeply for a domain.

The business use cases are situations where control outweighs raw capability. A hospital network that must keep patient data within its own environment can run a fine-tuned Gemma model on infrastructure it controls. A manufacturer can deploy a small Gemma model close to the factory floor where connectivity is limited. A startup can fine-tune Gemma into a cost-efficient specialist for one narrow task instead of paying for a flagship model on every request. The trade-off to remember: Gemma offers deployment freedom and customization; Gemini offers the strongest capability with no operational burden. When a scenario stresses open weights, self-hosting, strict data-locality needs, or deep customization on a budget, Gemma is the answer.

Imagen is Google's text-to-image foundation model: you describe an image in natural language and Imagen generates it, with strengths in photorealism, faithfulness to the prompt, and support for editing existing images, such as changing a background or extending a scene. Because the interface is plain language, non-designers can produce usable visuals, and design teams can produce far more options in far less time.

The business use cases center on visual content at scale. Marketing teams generate campaign imagery and seasonal variations without scheduling photo shoots. Retailers produce product visuals in different settings and styles. Product and UX teams turn written concepts into visual mockups during early exploration. Agencies iterate through dozens of creative directions in the time a single manual round used to take.

A concrete example: a furniture retailer preparing a holiday campaign needs its sofa photographed in a dozen differently styled living rooms. Traditionally that means staging, shooting, and editing each scene. With Imagen, the team generates the sofa in every setting from text descriptions, selects the strongest directions, and reserves the physical shoot for the one hero image. The strategic value is speed and volume of creative iteration, and, as with all generative media, businesses should pair Imagen with responsible-use review before publishing. When a scenario asks for still visual content, generated or edited from a description, Imagen is the answer.

Veo is Google's video generation foundation model: from a text description, or from a starting image, it produces high-definition video clips, with control over cinematic qualities like camera movement and visual style. Video is the most expensive and slowest content type for businesses to produce, which is exactly why a model that generates it from a prompt matters.

The business use cases mirror where video budgets go today. Marketing teams generate short-form video for ads and social channels without a production crew. Creative teams storyboard and pre-visualize concepts in motion before committing to a shoot, testing an idea in hours instead of weeks. Training and communications teams turn scripts into illustrative footage. E-commerce teams animate product concepts to show them in use.

A concrete example: a travel company wants ten short video variations promoting different destinations for a social campaign. Filming on location in ten countries is out of budget; with Veo, the team generates evocative destination clips from descriptions, tests which perform best with audiences, and invests real production money only in the winners. The distinction to hold onto for the exam is simple: Imagen produces still images, Veo produces moving video. When a scenario involves generating footage, animating a concept, or pre-visualizing a scene in motion, Veo is the answer.

Choose by asking two questions in order: what modality does the output require, and how much control does the business need. The first question usually settles it. If the task is to generate still images, use Imagen. If it is to generate video, use Veo. If it involves understanding, reasoning, conversation, or code, or mixed inputs like documents plus images plus audio, use Gemini.

The second question separates Gemini from Gemma for language tasks. Managed Gemini is the default: strongest capability, no infrastructure to operate, enterprise controls provided by the platform. Choose Gemma when the business specifically needs open weights, self-hosted or edge deployment, strict control over where data and the model live, or a deeply customized, cost-efficient specialist model.

Beyond modality and openness, the same selection factors apply that you would use for any foundation model decision: context window (how much material the model can consider at once), cost and performance for the expected volume, security and data-handling requirements, availability and reliability expectations, and how much fine-tuning or customization the use case demands. Note that the models are complements, not competitors: a single campaign might use Gemini to develop messaging, Imagen to create the imagery, and Veo to produce the video, and a business choosing Google Cloud gets all four through one platform rather than assembling separate vendors per modality.

A consumer-electronics company launching a smart-home device shows how the four models divide the work. The launch needs market analysis, a support experience, launch visuals, and promotional video.

The strategy team uses Gemini to analyze competitor materials, customer reviews of the previous model, and internal research documents, then drafts positioning and messaging from that grounded picture. The support organization builds an assistant, powered by Gemini as its reasoning engine, that answers setup questions by drawing on the product manual and troubleshooting guides.

The creative team uses Imagen to generate launch imagery, the device styled in modern kitchens, living rooms, and offices, iterating through directions before commissioning final photography, and uses Veo to produce short teaser clips showing the device in daily-life scenes for social channels. Meanwhile, the engineering team has a constraint: an on-device voice feature must work without sending audio to the cloud. They fine-tune Gemma into a compact model that runs on the device itself, keeping user audio local.

One launch, four models, each chosen by modality and control needs: Gemini for reasoning and conversation, Imagen for images, Veo for video, and Gemma where open weights and on-device deployment were non-negotiable. That mapping, use case to model, is precisely the judgment the exam tests.