Skip to main content

A powerful step forward in performance, control, and simplicity.

Nutanix Enterprise AI 2.4 introduces meaningful improvements that make managing inference endpoints faster, smarter, and more flexible—across edge, on-prem, or cloud environments.

Review the detailed release notes before deploying any product updates.

🧠 Smarter Performance, Built-In

  • CPU and GPU Acceleration
    Choose between GPU or Intel®️ AMX-enabled CPU acceleration for optimal performance—tailored to your workload.

  • Expanded Model Support
    Pre-validated support for new Hugging Face and NVIDIA models including Llama 3, Mistral, Stable Diffusion, and more.

  • Now with NVIDIA H200 GPU support
    Tap into the next generation of high-performance AI acceleration.

🛠️️ A Simpler, Smoother Workflow

  • New Endpoint Creation Wizard
    A guided, intuitive flow for building endpoints—no YAML needed.

  • Create Endpoints from Imported Models
    Go from model import to deployment in a single click.

  • Auto Model Size Detection
    Model size is now automatically detected when importing from Hugging Face.

  • HTTP Proxy Support
    Easily connect to external services—even from air-gapped or dark sites.

🔐 More Control, More Confidence

  • Fine-Grained Access to Models
    Control exactly which models your users can import or upload—via catalog, direct URLs, or manual upload.

  • Search, Filter, and Discover
    Improved model and endpoint search with case-insensitive matching.

📊 Better Insights at a Glance

  • Enhanced Endpoint Metrics
    Get visibility into model performance with new metrics:

    • Time to First Token (TTFT)

    • Time Per Output Token (TPOT)

    • Output Tokens per Second

  • Kubernetes Monitoring Made Easy
    A unified view of your cluster, nodes, and usage—all in one place.

  • Click-Through Model Links
    Quickly access Hugging Face or NIM model source pages right from your dashboard.

🧰 Support Gets Smarter

  • Support Bundles On Demand
    Generate support bundles from your Kubernetes cluster to streamline troubleshooting with Nutanix Support.

Minimum Requirements

  • Kubernetes 1.31, NVIDIA GPU Operator v25.3.0, NVIDIA Toolkit v1.17.5, and more.

  • Supported GPUs: NVIDIA L40S, A100, H100, H100-NVL, and H200.

For full software and GPU requirements, see the Nutanix Enterprise AI Guide.

🎯 Final Thought

With this release, Nutanix Enterprise AI becomes even more robust—giving enterprises the tools they need to securely deploy and monitor GenAI workloads at scale, with greater control and visibility than ever before.

Continue the conversation on the Nutanix Cloud Platform for AI forum. 

Hey Community!, Product Marketing here for Nutanix Enterprise AI (NAI) - happy to answer any questions.🤗

 


If I could highlight a key 2.4 component it would be the inclusion of the new open-weight, OpenAI models, GPT-OSS! A first from OpenAI, the creator of ChatGPT, these models show how OpenAI is shifting their strategic focus to private and sovereign AI.

NAI 2.4 includes pre-validated NVIDIA NIM microservice day 0 support for the release of OpenAI OSS models, gpt-oss-120b, and gpt-oss-20b. These are the first open-weight language models from OpenAI, the owner of ChatGPT. This moment is a seismic shift for proprietary AI offerings, where on-premises versions of their models are now being offered.

 

NAI leverages the NVIDIA NIM microservices to use these, which can be used in NAI by heading over to NVIDIA and signing up for a developer account (free!). NAI uses that account to create an API token to access the models (not recommended for production 😅 )

 

 

One more thing!

💃🏿🕺🏽💃NAI is now available as a 60-day trial/POC for ANY current Nutanix customer! 💃🏿🕺🏽💃Simply logon to the portal to start the request process: https://portal.nutanix.com/page/home

 

📢If required, reach out to me, mike.barmonde@nutanix.com for help!📢

 

Terms & Conditions