Can I keep using GPT models on Omnia?

Yes. Omnia implements OpenAI’s protocol natively, and Azure OpenAI through Azure AI Foundry.

What does moving cost?

Omnia is free while you build and the core is free in production. The enterprise licence is one price per environment per year from £30,000, with nothing metered.

omnia vs openai agent builder

Moving off Agent Builder? Omnia keeps the agent, its prompts and its tests in your own repository.

OpenAI’s deprecations page schedules Agent Builder, the Evals platform and reusable prompts to shut down on the same day, 30 November 2026. The pack Omnia keeps them in runs in your cluster, on GPT or any other model.

1 December 2026. A support agent built in Agent Builder last year.

on omnia

The agent is still answering. Its prompts, tools, policy and scenarios are a pack in your repository, and last year’s tests still run against it. docs

on openai agent builder

Agent Builder is scheduled to shut down on 30 November 2026, with the Evals platform and reusable prompts the same day. OpenAI suggests moving prompt content into application code and evals to Promptfoo. their docs

The agent and its tests live in your repository

Prompts, tools, policy and scenarios are one versioned pack, synced from git, OCI or a configMap. docs

OpenAI Agent Builder: Reusable prompts shut down on 30 November 2026; OpenAI’s advice is to move their content into your application code. source

Keep the model, or change it

Omnia speaks OpenAI’s protocol natively, alongside Anthropic, Gemini, Bedrock, Vertex, Ollama and vLLM. You can stay on GPT and test a second model against the same scenarios. docs

OpenAI Agent Builder: The suggested replacements are the Agents SDK or ChatGPT Workspace Agents. source

Evaluation that gates the release

Each version goes through its real tools with simulated users, voice and adversarial callers among them, and the release waits on the result. A conversation that fails in production becomes a scenario every later version has to pass. docs enterprise

OpenAI Agent Builder: The Evals platform shuts down on 30 November 2026, with a migration guide to Promptfoo. source

At a glance

Topic	Omnia	OpenAI Agent Builder, per its docs
Agent builder available after 30 November	yes The Arena builder writes the agent and its tests from a brief. docs enterprise	not documented Shuts down. source
Evals	yes Simulated users, voice, adversarial; gates the release. docs	not documented Shuts down; migrate to Promptfoo. source
Stored prompts	yes In a versioned pack in your repository. docs	not documented Shut down; move into application code. source
OpenAI models	yes Native OpenAI protocol, plus six others. docs	documented Yes. source
Runs in your infrastructure	yes Your Kubernetes cluster, or AgentCore, Vertex or Foundry. docs	not documented OpenAI-hosted. source

Common questions

When does Agent Builder shut down?: OpenAI’s deprecations page schedules Agent Builder, the Evals platform and reusable prompts to shut down on 30 November 2026.
Can I keep using GPT models on Omnia?: Yes. Omnia implements OpenAI’s protocol natively, and Azure OpenAI through Azure AI Foundry.
What does moving cost?: Omnia is free while you build and the core is free in production. The enterprise licencelicense is one price per environment per year from £30,000, with nothing metered.

Put your own agent through the gate.

Install Omnia in your cluster and run it free while you build.

OpenAI Agent Builder claims are from OpenAI’s published documentation, checked 22 October 2026. Spot something out of date? Tell us.

Install it in your cluster Or walk through it with us →