Red Hat AI Inference 3.5

Release notes

Highlights of what is new and what has changed with this Red Hat AI Inference release

Getting started

Getting started with Red Hat AI Inference

Distributed Inference with llm-d

Architecture, components, and deployment of Distributed Inference with llm-d for scalable LLM serving on Kubernetes

This content is not included.

Validated models

Red Hat AI validated models

This content is not included.

Supported product and hardware configurations

Supported product and hardware configurations for deploying Red Hat AI software

Deploy the standalone Red Hat AI Inference container in OpenShift Container Platform

Deploy the standalone Red Hat AI Inference container in OpenShift Container Platform clusters that have supported AI accelerators installed

Deploy the standalone Red Hat AI Inference container in a disconnected environment

Deploy Red Hat AI Inference in a disconnected environment using OpenShift Container Platform and a disconnected mirror image registry

Inference serving language models in OCI-compliant model containers

Inferencing OCI-compliant models in Red Hat AI Inference

Speculative decoding

Speculative decoding with Red Hat AI Inference

Inference serving Mistral 3 models

Inference serving Mistral 3 models with Red Hat AI Inference

Inference serving geospatial foundation models

Inference serving geospatial foundation models with Red Hat AI Inference

Red Hat AI Model Optimization Toolkit

Compressing large language models with the LLM Compressor library

vLLM server arguments

Server arguments for running Red Hat AI Inference

Extending Red Hat AI Inference with tool calling capabilities

Configuring tool calling and chat templates for AI Inference

Deploy Distributed Inference with llm-d on OpenShift Container Platform

Deploy and serve large language models at scale on OpenShift Container Platform

Deploy Distributed Inference with llm-d on AWS, Azure, or CoreWeave Kubernetes Service

Deploy Distributed Inference with llm-d on Azure or CoreWeave Kubernetes Service

Monitor and troubleshoot Distributed Inference with llm-d deployments

Monitor and troubleshoot Distributed Inference with llm-d deployments

Batch process inference requests with Distributed Inference with llm-d

Batch process inference requests with Distributed Inference with llm-d

Deploy large MoE models across multiple GPUs with WideEP

Deploy Distributed Inference with llm-d on Azure or CoreWeave Kubernetes Service

Deploy Models as a Service on non-OpenShift Kubernetes

Deploy and configure Models as a Service on non-OpenShift Kubernetes environments

Manage mixed workloads by using priority queuing

Manage mixed workloads by using priority queuing

Tool calling for Distributed Inference with llm-d deployments

Tool calling for Distributed Inference with llm-d deployments

Upgrade Distributed Inference with llm-d on managed Kubernetes

Upgrade the Distributed Inference with llm-d infrastructure stack on managed Kubernetes clusters

This content is not included.

Red Hat Enterprise Linux AI

Switch to the Red Hat Enterprise Linux AI documentation

This content is not included.

Red Hat OpenShift AI

Switch to the Red Hat OpenShift AI documentation

This content is not included.

Red Hat AI Enterprise

Switch to the Red Hat AI Enterprise documentation

Red Hat AI Inference Server 3.3

Switch to the Red Hat AI Inference Server 3.3 documentation

This content is not included.

Red Hat AI Foundations

Explore no-cost courses to boost your AI knowledge and get hands-on experience with Red Hat AI products while earning a certificate

This content is not included.

Red Hat AI learning hub

Explore the curated set of third‑party models validated for Red Hat AI products, ready for fast, reliable deployment