Red Hat Released AI 3.5 With Expanded Operational Tools
The latest release brings enhanced observability, multi-tenancy, and compliance features to enterprise AI infrastructure.
Updated on Sept. 29, 2026 in Artificial Intelligence
Live Poll
Do you trust that businesses have the necessary controls to deploy AI safely at scale?
Red Hat has released Red Hat AI 3.5, a version update that introduces new capabilities for safety, observability, multi-tenancy, and agentic application development. The update is designed to help organizations transition AI workloads from initial pilot phases into scalable operations.
Why it matters
As enterprise AI moves past early experimentation, companies require increased operational rigor to manage deployments. Red Hat AI 3.5 aims to address this need by providing tools for model verification and infrastructure management on shared hardware.
The release introduces version 3.5, featuring Inference-Time Scaling that adjusts compute resources based on query complexity. It also integrates EvalHub for model verification and regulatory compliance, alongside observability dashboards for tracking GPU utilization and inference health.
The players
Red Hat
An enterprise software company specializing in open-source solutions, including the OpenShift container orchestration platform.
The details
Red Hat AI 3.5 leverages hosted control planes on Red Hat OpenShift Virtualization to provide hardware-to-software isolation, which allows for multi-tenancy on shared GPU infrastructure. Developers can build agentic applications—systems capable of autonomous decision-making—using visual pipelines and AutoRAG, a tool for retrieval-augmented generation that now supports multilingual document processing and conversational testing.
Timeline
Red Hat released Red Hat AI 3.5 on September 29, 2026.
The Tech Race
Red Hat is positioning its software stack to compete in the enterprise AI infrastructure market by prioritizing operational stability and regulatory compliance. This release follows a broader trend where incumbent cloud and virtualization vendors are adapting their platforms to handle the complex requirements of scaled generative AI.
Enterprise teams can now utilize visual pipelines to configure RAG applications, potentially reducing the development time required for complex AI workflows. The new observability tools are currently available for users running models on Red Hat OpenShift infrastructure.
The takeaway
The trajectory for enterprise AI is shifting toward managing infrastructure complexity and model reliability rather than just model training. Technical teams should watch for performance benchmarks comparing the new multi-tenancy features against legacy dedicated-GPU deployments to assess scalability efficiency.
Further reading
For more on the evolving enterprise software landscape, browse Artificial Intelligence.
Source note: This article includes information reported by IT-Online.
Live Poll
Do you trust that businesses have the necessary controls to deploy AI safely at scale?
