📼 Rewatch VS Code Live at MS Build 2026

Any harness, any model

Choose the harness you want the agent to run in, then pair that agent with the model that fits the task.

Download Compare agent harnesses

A setup that adapts to the task

Keep the harness you know, pick the model that fits the task, and decide where that model runs.

Work in a familiar harness

Use the agent harness you already know, such as GitHub Copilot, Claude Code, or Codex.

Pick the best model for the task

Choose from multiple providers, including Anthropic, OpenAI, Google, and Microsoft, plus open weight models, and match speed, reasoning, and context window to the work at hand.

Choose where your model runs

Bring your own key to connect a provider account directly, route through infrastructure you control, or run a model locally to work offline or keep code on your machine.

Start with the harness that fits

The agent harness determines where work runs, how it interacts with your workspace, and when it asks for approval. Switch harnesses mid-session and VS Code carries your conversation history and context along with it.

Learn about code isolation and handoff
Agent picker in VS Code with Copilot, Codex, and Claude available. Agent picker in VS Code with Copilot, Codex, and Claude available.
Language Models editor in VS Code showing available models and their capabilities. Language Models editor in VS Code showing available models and their capabilities.

Pair it with the right model

Switch models as your needs change, compare how they handle the same task, and adopt new releases as they land.

See supported language models

Get more from every AI credit

Different models consume AI credits at different rates, so the model you pick shapes what a task costs.

Optimize AI credit usage
Agent session for applying accessibility improvements to a web page, showing changes to the code and option to create a PR. Agent session for applying accessibility improvements to a web page, showing changes to the code and option to create a PR.

Bring your own key

Add your own API key to use a provider account directly, or configure a custom endpoint to route requests through infrastructure you control.

Configure your own model provider
Model picker in VS Code showing models from multiple providers. Model picker in VS Code showing models from multiple providers.

And much more

Language models

See which models are available, how to add a provider, and how to switch between them.

Auto model selection

Let VS Code route each request to an efficient model based on task complexity and service health.

Configure thinking effort

Dial reasoning up for complex problems, or down to reduce latency and AI credit use.

Cache Explorer

Diagnose prompt cache misses between requests to reduce token cost and latency.

Custom agents

Pin a specific model to a purpose-built agent so the right one is always used for the job.

Pick the agent and model that fit the work

Bring the agents and models you already use, then switch either one as your needs change.

Download VS Code Explore language models