
When you're getting started with an AI project, choosing the right model is only half the battle. You also need to configure environments and install dependencies while connecting the tools that make your workflows run smoothly. Setting things up can quickly become a bottleneck.

Prebuilt templates simplify the process – and **Vast.ai** gives you plenty to choose from!

Our [Model Library](https://vast.ai/model-library) offers a model-first way to deploy AI, with templates that launch ready-to-run environments for some of the most-used AI tools today. Vast.ai also offers [lower-level templates](https://cloud.vast.ai/templates) for frameworks and inference engines like PyTorch and vLLM for more customizable setups.

Each pre-configured template is organized by modality and use case so you can quickly find what you need – and you retain full custom control over the underlying runtime, parameters, and infrastructure.

### Popular Templates in the Vast.ai Model Library

From image, text, and audio-video generation to multimodal AI workflows, the following are a few of the most popular templates being deployed on Vast.ai right now.

#### 1. Flux.2-Dev – Rectified Flow Image Generation

[**FLUX.2-Dev**](https://vast.ai/model/flux.2-dev) is a 32-billion parameter rectified flow transformer built for advanced text-to-image generation and editing. The model processes text prompts alongside optional reference images, so you can modify existing images or combine multiple visual references without additional fine-tuning.

Its ability to maintain consistent characters, objects, and styles across generations makes it ideal for character design, visual storytelling with scene continuity, digital art with precise style control, and design prototyping.

#### 2. LTX-2 – Native Audio-Video Generation in a Single Workflow

[**LTX-2**](https://vast.ai/model/ltx-2) is a 19-billion parameter Diffusion Transformer (DiT) model that generates audio and video together in a single pass. It creates synchronized audiovisual output from text prompts, images, and other inputs – keeping dialogue, motion, and ambient sound aligned throughout the clip.

One of LTX-2's biggest advantages is its light and iterative workflow. Since the model generates video in a compressed latent space before converting it to full resolution, it supports faster iteration and more efficient memory usage than many traditional video models.

#### 3. WAN 2.2 – Efficient Text-to-Video and Image-to-Video Generation

[**WAN 2.2**](https://vast.ai/article/wan-2-2-explained-new-approach-ai-video-generation) is a video generation model that produces short cinematic clips from either text prompts or static images. What sets it apart is its Mixture-of-Experts (MoE) architecture. The model switches between specialized experts that handle early scene structure and later visual refinement. This results in smoother motion and higher visual fidelity along with more efficient use of compute.

Available on Vast.ai in both text-to-video ([T2V](https://vast.ai/model/wan-2.2-t2v-a14b)) and image-to-video ([I2V](https://vast.ai/model/wan-2.2-i2v-a14b)) variants – with the latter offering optional guidance via text prompts – WAN 2.2 gives you fine control over motion, lighting, and scene composition. Popular use cases include rapid prototyping of video concepts based on text descriptions and product animations generated from still images.

> Not sure whether **WAN 2.2** or **LTX-2** fits your workflow best? Our detailed [comparison post](https://vast.ai/article/wan-2-2-vs-ltx-2-which-ai-video-model-should-you-use) explores the strengths of each model and which one to choose for your specific needs!

#### 4. ACE-Step V1 – Fast Text-to-Music Generation

[**ACE-Step V1**](https://vast.ai/model/ace-step-v1-3.5b) is an open-source music generation model that creates full musical compositions from text prompts in **17 languages**. Thanks to its diffusion-based architecture and efficient audio compression pipeline, it generates music about **15x faster** than comparable alternatives.

Musical quality is high, with controllable duration (up to five minutes works best) and coherent output across melody, harmony, and rhythm. You can even edit lyrics and manipulate vocals. Overall, ACE-Step V1 is a practical tool for fast experimentation with musical ideas, as well as voice cloning applications and music remixing and style transfer.

#### 5. Dia 1.6B – Realistic Dialogue Generation from Text

[**Dia 1.6B**](https://vast.ai/model/dia-1.6b) is a text-to-speech model that produces highly realistic dialogue in English from given transcripts. Using [S1] and [S2] speaker tags, it creates multi-speaker conversations that can also include nonverbal expressions like laughter, coughing, and gasps.

Audio conditioning enables precise control over emotion and tone, and the model even supports voice cloning functionality. Dia 1.6B's applications include conversational audio content, accessibility tools, and voice synthesis applications that require consistent, natural-sounding speech.

#### 6. DeepSeek V3.2 Exp – Long-Context Sparse Attention LLM

[**DeepSeek V3.2 Exp**](https://vast.ai/model/deepseek-v3.2-exp) is an open-source experimental large language model designed for long-context reasoning and advanced text generation. It introduces DeepSeek Sparse Attention, a novel mechanism that enables efficient processing of extended documents and conversations without sacrificing output quality.

The model is well suited for long-form document analysis and generation, research that requires extended reasoning, code generation and debugging, and even customer support with context-aware responses. Notably, it has a license that offers full commercial use permissions.

#### 7. Qwen3.5 397B A17B – Multimodal Reasoning and Agentic Workflows

[**Qwen3.5 397B A17B**](https://vast.ai/model/qwen35-397b-a17b) is the open-weight of the first model in the new Qwen3.5 series. It's a frontier-level native vision-language model with outstanding results across numerous benchmarks.

Built with a sparse mixture-of-experts (MoE) architecture and hybrid attention mechanisms, Qwen3.5 397B A17B supports highly complex tasks in areas like mathematical reasoning and problem solving, multi-turn agentic workflows, code generation and debugging, document analysis, video understanding, visual question answering, and multilingual interaction across **200+** languages and dialects.

### Launch Templates in Minutes on Vast.ai

Today's AI workloads span an incredibly wide range of modalities. The templates above represent just a small snapshot of what developers are building with on Vast.ai right now.

With the **Vast.ai** [Model Library](https://vast.ai/model-library), getting started on your own AI projects is easy. Simply browse templates by use case, select the recommended GPU hardware, and launch a production-ready environment in minutes!


Discover the most popular prebuilt templates on Vast.ai for image, video, audio, and text AI workloads — from Flux.2-Dev to Qwen3.5.

7 Popular Model Templates for AI Workloads on Vast.ai Right Now


As 2025 comes to a close, we're looking back at the milestones and updates that shaped the past year for us here at Vast.ai.

What a year it's been! From new hardware to major security achievements and product launches, here are our top moments of 2025.

## Expanding Our Global GPU Fleet

The past year saw some of the biggest hardware expansions in Vast.ai's history.

In 2025, we added NVIDIA's latest RTX 5000-series GPUs – including the RTX 5090 – followed by high-performance machines like the RTX PRO 6000 WS and the H200 and B200. These additions supported higher-throughput workloads and more efficient AI inference across a [wide range](https://cloud.vast.ai/) of hardware profiles.

Another highlight: our RTX 5090 inventory surpassed 1,000 machines this year!

## Security and Compliance Milestones

This year, Vast.ai achieved both **SOC 2 Type I** and **Type II** certification. These independent audits validated that the security infrastructure we've built over 6+ years of trusted service continues to meet rigorous compliance standards.

We take this responsibility seriously. Data privacy and integrity remain core priorities in everything we build. As we continue to advance our [security and compliance](https://vast.ai/compliance) roadmap, we're committed to upholding the trust you place in us.

## New Templates and Tools

We introduced a number of ready-to-use templates this year, to help you move from idea to execution quickly and seamlessly. Some notable ones included:

- [Langflow + Ollama](https://docs.vast.ai/langflow-ollama) for visual AI workflow design
- [Ostris AI Toolkit](https://cloud.vast.ai/?ref_id=62897&template_id=81e6e9dd266d9bdbcbb0fd48b9afe357) for custom Stable Diffusion and LoRA training
- [Open-Sora](https://cloud.vast.ai/?ref_id=62897&creator_id=62897&name=Open-Sora) and [LTX Video](https://cloud.vast.ai/?ref_id=62897&creator_id=62897&name=ComfyUI%2520%252B%2520LTX%2520Video) (ComfyUI) for video generation workflows
- [ComfyUI ACE Step](https://cloud.vast.ai/?ref_id=62897&creator_id=62897&name=ACE%2520Step%2520Music%2520Generation) for AI-assisted music creation
- [InvokeAI](https://cloud.vast.ai/?ref_id=62897&template_id=c309df384010cf9d3dc9abb9e0a5a6fd) for image generation in a professional web interface
- [Dia 1.6b TTS](https://cloud.vast.ai/?ref_id=62897&template_id=6ba071acb5f590913c5eaf6c2dc3d0db) for lifelike text-to-speech

But the biggest update of the year was still ahead...

## Introducing Vast.ai Serverless

The headline story of 2025 was the [recent launch](https://vast.ai/article/vast-ai-serverless-automated-gpu-scaling) of **Vast.ai Serverless** – a fully automated GPU compute layer designed for AI inference at scale.

With our Serverless offering, you can easily run workloads without managing instances on your own or overpaying for capacity. You simply send inference requests and the platform automatically selects the right hardware, routes workloads across regions as needed, and predictively scales up and down based on usage.

In short, Vast.ai Serverless gives you access to a diverse, globally distributed GPU fleet with a serverless model that applies predictive optimization and flexible scaling – with none of the overhead. We're thrilled to finally make truly serverless AI inference a reality on Vast.ai!

## Looking Ahead

To all of our users and the wider Vast.ai community, thank you for being part of our journey this year. It's been an incredible year of growth, and we're grateful to everyone building and innovating with us.

Here's to an even stronger year ahead!


As 2025 comes to a close, we're looking back at the milestones and updates that shaped the past year for us here at Vast.ai.

Vast.ai's 2025 Year in Review


# Running ComfyUI on Vast.ai Serverless: Simple, Scalable Image Generation

ComfyUI has become one of the most popular open-source AI tools for image generation, for good reason.

Rather than typing a prompt and hoping for the best, you build the entire process yourself through a **node-based interface that gives you full, visual control** over how images are created. Every part of the workflow – from models, prompts, and parameters – can be adjusted or branched however you like, and workflows themselves are also shareable and reusable.

While ComfyUI runs well locally for smaller projects, it's a whole other challenge to use it for heavier custom workflows or at production scale. Complex ComfyUI graphs can exceed consumer GPU memory or demand more throughput than a single machine can handle – or simply take too long to iterate on locally.

That's where **Vast.ai Serverless** comes in.

### ComfyUI Meets Vast.ai Serverless

Vast.ai Serverless makes it easy to run your ComfyUI workflows at scale, thanks to a ready-to-use template that handles all of the setup for you – and no need to manage GPUs or monitor capacity.

Simply send requests to ComfyUI and have your generated assets automatically uploaded to S3-compatible storage. The template will return pre-signed URLs in response to requests, as well as detailed process updates while the workflow runs.

It's a straightforward way to integrate ComfyUI into your applications or pipelines without worrying about infrastructure overhead.

### What Comes in the ComfyUI Serverless Template

Our ComfyUI Serverless template provides [everything you need](https://www.google.com/url?q=https://docs.vast.ai/documentation/serverless/comfy-ui&sa=D&source=editors&ust=1766188232265630&usg=AOvVaw0k4v0rLcf_0MltEhly3_ZW) right out of the box:

- **ComfyUI** and its **API wrapper** (which manages asset uploads)
- **Stable Diffusion 1.5** for benchmarking
- **PyWorker** for processing JSON workflows
- **Provisioning script** for custom configuration

The template saves you from building or maintaining your own Docker image and ensures you get a consistent environment every time a worker spins up. Once the template is in place, several capabilities come built into the workflow by default.

#### 1. Built-In Asset Handling with S3 Storage

Generated assets are automatically uploaded to your private S3-compatible storage buckets, and pre-signed URLs provide secure, temporary access to download results directly. This keeps storage and retrieval simple, even as workloads scale.

Just add your credentials once in your account settings, and every worker can use them. If you need different buckets for different jobs, you can override these variables on a per-request basis.

#### 2. Flexible Configuration When You Need It

With ComfyUI Serverless, you can keep things simple or tailor your setup with optional configurations. For instance:

- Add a webhook to be notified after generation completion or failure
- Use a custom PyWorker git repository URL
- Adjust benchmark settings to better reflect your workload

With these configuration tools, you're free to adapt the template to your unique needs.

#### 3. Distributing Work Through Benchmarking

Each worker runs a quick benchmark when it initializes, to validate GPU performance and calculate a performance score. This score is what determines how requests are distributed across workers.

The template includes a default Stable Diffusion 1.5 benchmark for this purpose, but you can configure it to match your actual workload complexity or even supply your own custom benchmark.

### Try It Out Interactively Before Going Serverless

The ComfyUI Serverless template is designed for serverless execution, of course, but you can test it as an interactive instance first. Interactive testing helps confirm that everything works as expected before you start sending requests through Serverless.

You'll have to enable port forwarding to access applications over the public interface, but the benchmarking process will be visible in the instance logs either way.

### A Smoother Path to Production-Scale ComfyUI

Vast.ai Serverless and our ComfyUI template lets you go from a local workflow to a scalable, on-demand system with minimal setup. The environment is preconfigured – with customizable adjustments as needed – and assets are handled automatically while workloads are routed to the right GPUs based on real performance.

For creative work beyond images, we offer Serverless templates for [text-to-video](https://www.google.com/url?q=https://docs.vast.ai/documentation/serverless/comfyui-wan-2.2&sa=D&source=editors&ust=1766188232275066&usg=AOvVaw33TRQTBXX80tZ040XucSdk) and [text-to-music](https://www.google.com/url?q=https://docs.vast.ai/documentation/serverless/comfyui-acestep&sa=D&source=editors&ust=1766188232275242&usg=AOvVaw3Dnzq-bbQYJjA76I3vHcU3) workflows, as well.

Ready to see ComfyUI in action? Just launch the [ComfyUI Serverless template](https://www.google.com/url?q=https://cloud.vast.ai/?ref_id%3D62897%26creator_id%3D62897%26name%3DComfyUI%2520(Serverless)&sa=D&source=editors&ust=1766188232275749&usg=AOvVaw0clIXR2u6vdhxqRdqoPmHE) on Vast.ai and start running workflows at scale. It's that easy.


ComfyUI has become one of the most popular open-source AI tools for image generation, for good reason.

Running ComfyUI on Vast.ai Serverless: Simple, Scalable Image Generation


[Video generation with AI](https://docs.vast.ai/video-generation) has come a long way. If you saw the infamous AI-generated video of Will Smith eating spaghetti a couple of years ago, you probably remember how unnatural and distorted the clip looked\! But that was just the beginning.

Things look very different today, with more advanced video generation tools producing much higher quality results. Today's models can create photorealistic visuals with smooth, consistent motion as well as stylized and cinematic effects if that's what's called for.

AI video generation has become a practical tool for anyone who needs an easy way to go from text to video, no matter the project. The right templates make that process even simpler.

With that in mind, let's take a look at a few of the best [AI video generation](https://docs.vast.ai/video-generation) templates available today.

### 1\. [ComfyUI \+ LTX Video](https://cloud.vast.ai/?ref_id=62897&creator_id=62897&name=ComfyUI%20%2B%20LTX%20Video)

[**LTX Video**](https://github.com/Lightricks/ComfyUI-LTXVideo) is a group of open-source AI video models built for high-quality diffusion-based video generation. It offers both speed and precision, generating high-resolution videos **up to 60 seconds long** (with some versions of the model limited to **10 seconds**), often in less time than it takes to watch the resulting video.

How is this possible? LTX Video uses a unique approach where it **compresses video data 192 times smaller than the original** – past the limits of what most models can achieve – while still maintaining crisp visuals and smooth, realistic motion. A bonus is that this drastically reduces the amount of memory and compute required. In fact, LTX Video is optimized specifically for consumer-grade GPUs.

While most diffusion models use a two-step pipeline where a second-stage model sharpens details, LTX Video does decoding and cleanup in just one step. With a Video-VAE (variational autoencoder) and a denoising transformer combined into a single integrated system, there's no need to patchify the video stream, and high-quality results are possible even under heavy compression.

This [template](https://cloud.vast.ai/?ref_id=62897&creator_id=62897&name=ComfyUI%20%2B%20LTX%20Video) pairs LTX Video with **ComfyUI**, a modular, node-based interface for Stable Diffusion workflows. Together, they serve as a seamless way to run efficient, high-compression video pipelines with a user-friendly interface.

### 2\. [Mochi](https://vast.ai/article/generating-videos-with-mochi-on-vast.ai)

Designed to excel in quality of motion as well as prompt adherence, [**Mochi**](https://github.com/genmoai/mochi) is a state-of-the-art open-source video generation model that performs competitively with leading closed models.

It demonstrates strong motion quality, generating smooth videos at **30 frames per second for durations of up to 5.4 seconds.** Fluid dynamics, fur and hair simulation, and human movement are all rendered with impressive realism.

Mochi's alignment with text prompts is also exceptional. Generated videos accurately depict the given input, allowing users to control characters, settings, and actions with a high degree of detail. Prompt adherence is benchmarked with a vision-language model and generated videos are evaluated using Gemini to ensure accurate, consistent results.

However, keep in mind that the initial release of Mochi only generates videos at **480p**, and the model is optimized for photorealistic styles, so animated content is not recommended.

### 3\. [Open-Sora](https://cloud.vast.ai/?ref_id=62897&creator_id=62897&name=Open-Sora)

[**Open-Sora**](https://github.com/hpcaitech/Open-Sora) is an open-source video generation model with a streamlined, user-friendly platform that lowers the barrier to efficient, high-quality video content creation. It can produce videos **up to 15 seconds long, at resolutions of up to 720p, with any aspect ratio.**

The model also supports a wide range of visual generation tasks: text-to-image, text-to-video, and image-to-video generation. This flexibility means users can run diverse workflows within a single framework.

Notably, Open-Sora introduces two innovations. The first is the Spatial-Temporal Diffusion Transformer (STDiT), a framework that decouples spatial and temporal attention. The second is a highly compressive 3D autoencoder that not only makes representations more compact but also accelerates training with its own ad hoc training strategy.

These features allow Open-Sora to generate high-quality results more efficiently – and with a ready-to-use template, it's easier than ever to start experimenting.

### Final Thoughts

AI video generation has matured rapidly in just a few short years. Instead of glitchy experiments, we now have reliable tools that deliver polished video content.

At **Vast.ai**, we believe in democratizing [access to these advanced AI tools](https://docs.vast.ai/video-generation), making them easy to use and available to everyone without the overhead of complex setup.

With our pre-built templates for [LTX Video](https://cloud.vast.ai/?ref_id=62897&creator_id=62897&name=ComfyUI%20%2B%20LTX%20Video), [Mochi](https://vast.ai/article/generating-videos-with-mochi-on-vast.ai), and [Open-Sora](https://cloud.vast.ai/?ref_id=62897&creator_id=62897&name=Open-Sora), you can launch cutting-edge video generation models in minutes – backed by affordable, on-demand cloud GPUs that save you from having to purchase and maintain your own expensive hardware. Use only the compute you need, when you need it.

Ready to experiment? Explore the [video generation templates](https://docs.vast.ai/video-generation) on Vast.ai and bring your ideas to life today\!


Generate AI videos with LTX Video, Mochi, and Open-Sora on Vast.ai. Step-by-step templates for high-quality, photorealistic text-to-video workflows with affordable GPU hosting.

Vast.ai: Best Video Generation Templates 2025


Imagine you are working on a data-intensive machine learning project and require a Linux environment optimized for deep learning frameworks. How can you easily and efficiently obtain one and still enjoy endless customizability? We recommend using templates to create Linux Docker instances with predefined settings.

This is where Vast.ai comes in. We provide containerized Linux Docker instances for rent along with a wide range of [templates](https://cloud.vast.ai/templates/) to suit numerous use cases for our customers.

## Templates – the Building Blocks of Customization

Think of a template as a blueprint for your instance. Each template is a saved set of initialization information that specifies what Docker image to load and includes settings that can only be configured when the instance is created.

So in the example project above, where you would like to make use of a deep learning framework, you could select a PyTorch or TensorFlow template on Vast. Then, with just a few clicks, you can create a new instance with your preferred framework, preset environment variables, and networking configurations ready to go.

With Vast templates, you have a multitude of these powerful tools in an easy-to-use format at your fingertips – from AI-powered image generation with [ComfyUI](https://vast.ai/article/getting-started-with-comfy-UI) to running large language models (LLM) via the pre-built [Oobabooga](https://vast.ai/article/deploying-oobabooga-for-running-large-language-models) template, and much more.

## How to Use Templates on Vast.ai

To get started, access the primary instance configuration menu by clicking on the [Templates link](https://cloud.vast.ai/templates/) in the navigation panel on the Vast console.

There are three different template categories for you to choose from:

- **Recommended** – templates that will load a smart set of options for typical use cases.

- **Popular** – a list of the current most popular templates used on Vast.

- **Recent** – where your customized and selected templates are saved.

To use any template as is, simply click “Select” to pick your desired template. After that, you will be directed to the search interface where you can find a suitable GPU offer and move forward from there!

If you prefer to customize a template for your own specific needs, click “Edit” to bring up the template editor. Editing any template creates a new version of that template, which will then appear in your Recent section after being saved.

You can also customize your template within the search interface via a separate configuration menu, although the settings are the same either way. For more information about template customization and selecting launch mode and disk size, see our detailed template guide [here](https://vast.ai/docs/instance-setup/templates).

## Making the Most of Your Templates

Vast also offers a [referral reward program](https://vast.ai/docs/account/referral-program). You can use a simple referral link, found on your account page, or drive traffic to a specific public template that you have created. Here’s how template referral works:

1. Make your own Vast template configured for the use case you want to promote and set the template to public.

2. Refer people to your template link (found next to the name of the template), which will autoload the Docker image and predefined settings for your use case, making it much easier for new users to get started on Vast.

3. Earn 3% of the credits purchased by anyone who registers a new client account via your public template referral link – and continue to earn 3% for the lifetime of the account!

4. Keep your referral credits as a positive balance in your account for your own use on Vast, or convert up to 75% of referred credits into cash that can be withdrawn via Stripe Connect, PayPal, or Wipe.

For the nitty-gritty details about our referral program, check out the full guide [here](https://vast.ai/docs/account/referral-program).

## Conclusion

With the help of our easy-to-use templates (as well as our referral program, should you choose to make use of it), we hope you will discover an even more accessible and cost-effective way to leverage the power of GPU rentals. At Vast.ai, we’re proud to be able to bring you a world of resources while minimizing the barriers of cost and complexity.

If you have any questions or would like further assistance, please feel free to contact us via the chat box below, or drop us a line at [support@vast.ai](mailto:support@vast.ai). We look forward to hearing from you.


Posts about: Templates

7 Popular Model Templates for AI Workloads on Vast.ai Right Now

Vast.ai's 2025 Year in Review

Running ComfyUI on Vast.ai Serverless: Simple, Scalable Image Generation

Vast.ai: Best Video Generation Templates 2025

Unlocking the Power of templates - Linux Docker Instances