Hyperstack Weekly Rundown 43: Simplified Kubernetes Deployment & Smarter Volume Management 🚀

We’re back with your weekly dose of Hyperstack updates!

Grab your coffee and stretch out that scroll finger, here’s what’s new this week. From kubernetes updates to tutorials and exciting blogs, there’s plenty to explore.

Let’s jump in! 

What's New on Hyperstack

Here's what's new on Hyperstack this week:

New Kubernetes Versions Supported

You can now launch clusters with Kubernetes 1.27.8, 1.32.8 and 1.33.4. Choose your preferred Kubernetes version during cluster deployment, and Hyperstack will automatically provision it with the correct compatible image. This means less setup and more time building.

Smarter Cluster Deployment

We’ve simplified the cluster creation process. Instead of manually choosing an image, you only select the Kubernetes version and we’ll handle the rest. This reduces complexity and ensures your clusters are always deployed with the optimal configuration.

API Updates for Better Automation

• List Cluster Versions API now returns detailed metadata about each version, including supported regions and pre-configured images. This gives your DevOps team clearer visibility and control when automating deployments.
• Create Cluster API no longer requires an image name. Just provide the Kubernetes version and provisioning happens seamlessly.

Latest Fixes and Improvements

• Pagination on the Volumes Page makes it much easier to navigate large fleets of volumes.
• When a volume is protected, it now shows a lock icon with a tooltip on attached virtual machines. You can view the protection status directly through the Volume Attachment APIs.
• Cluster creation and node/group scaling are now faster, so you can deploy and expand workloads with improved speed and efficiency.

Reserve NVIDIA RTX A6000 GPUs

Running 3D rendering, large-scale training cycles or graphics-intensive AI projects? The NVIDIA RTX A6000 is built for such demanding workloads while staying highly cost-effective on Hyperstack.

If your projects won’t slow down anytime soon, reserving capacity today on Hyperstack gives you a lower rate than our on-demand pricing. You can reserve the NVIDIA RTX A6000 GPUs for $0.35/hour. That means:

• Predictable performance for long projects
• No waiting on capacity when demand spikes
• Priority support to keep workflows smooth

Talk to Our Team to Reserve NVIDIA RTX A6000 GPUs →

New on our Blog

Check out our latest blogs on Hyperstack:

Deploying and Using Qwen3-Next-80B-A3B On Hyperstack: A Step-by-Step Guide

Qwen3-Next-80B-A3B is one of the latest models in the Qwen3-Next family, designed to combine massive model capacity with efficient inference. Built as a Mixture-of-Experts (MoE) model, it contains 80 billion parameters in total but only activates about 3 billion of them per token. This gives it the power of a very large model while keeping the speed and cost closer to that of a smaller one. In our latest tutorial, we explore how to deploy and use Qwen3-Next-80B-A3B on Hyperstack. From setting up your environment to running tasks, we guide you through each step to help you get started with Alibaba's latest model. 

Check out the full tutorial below!

What is AI-as-a-judge: And Why it Matters in LLM Evaluation

AI-as-a-Judge or LLM-as-a-Judge is the practice of using one AI model to evaluate the outputs of another AI model. Instead of relying entirely on human reviewers or rigid automated metrics, this approach allows an AI to act as a “QC”, analysing whether another AI’s response meets specific expectations. In our latest blog, we explored AI-as-a-Judge, a method where one AI evaluates another’s outputs, assessing qualities like tone, bias, and persona adherence. We discussed AI-as-a-Judge vs Human Evaluation, their pros and cons and showed how to implement custom evaluations in Hyperstack AI Studio.

Check out the full blog below!

How to Run the Flux Model on ComfyUI with NVIDIA H100 GPUs: A Quick Guide

Flux is one of the latest and most powerful text-to-image diffusion models, built to provide higher realism, sharper details and better prompt alignment compared to older models like SDXL. Unlike standard diffusion models, Flux integrates multi-modal conditioning (text, style prompts and additional context) to generate outputs that better align with user intent. In our latest tutorial, we walk you through how to run the Flux model on ComfyUI using NVIDIA H100 GPUs on Hyperstack. You’ll learn what Flux is, why ComfyUI is ideal for it and how H100 PCIe GPUs make it the best choice for smooth AI model deployment. 

Check out the full blog below!

Have an idea you'd like to see in Hyperstack? Let’s bring it to life.

At Hyperstack, we’re committed to continuous improvement and your ideas are a key driver of our innovation.

→ Is there a feature you’ve been waiting for?

→ Something that could speed up your workday?

→ Or a tweak that would make things feel effortless?

Tell us what would make your Hyperstack experience even better. Your feedback sets the direction for what we build next.

Share Feature Request →

That’s a wrap for this edition of the Hyperstack Weekly Rundown! Stay tuned for more updates, new features and announcements next week. Until then, don’t forget to subscribe to our newsletter for exclusive insights and the latest scoop on AI and GPUs - delivered right to your inbox!

Until next time,

The Hyperstack Team.