News

VS Code Update Helps Devs Select, Contribute AI Models

Microsoft announced Visual Studio Code 1.104, the August 2025 update, with new functionality for selecting and contributing AI models, security, productivity and more.

"This iteration, we're introducing auto model selection in chat," the dev team said today. "When you choose the Auto model in the model picker, VS Code automatically selects a model to ensure that you get the optimal performance and avoid rate limits."

This feature is currently in preview and Microsoft is rolling it out to all GitHub Copilot users in VS Code in the following weeks, starting with the individual Copilot plans.

Choosing Auto in the model picker lets VS Code select a model to optimize performance and help avoid rate limits. Auto currently chooses between Claude Sonnet 4, GPT-5, GPT-5 mini, GPT-4.1, and Gemini Pro 2.5, with visibility into the selected model on hover. VS Code also applies a variable model multiplier and a 10% request discount for paid users.

Auto Model Selection
[Click on image for larger view.] Auto Model Selection (source: Microsoft).

VS Code 1.04 also finalized an API that lets extensions contribute one or more language models. Users can pick contributed models in the Chat model picker, and Microsoft points to example extensions including AI Toolkit for VS Code, Cerebras Inference, and Hugging Face. Guidance for extension authors is provided in the Language Model Chat Provider docs and sample accessible from the announcement post.

Security: Confirm Sensitive File Edits and Safer Terminal Actions
Agent mode now asks for confirmation before modifying sensitive files, with defaults covering common system folders, dotfiles, and locations outside the workspace. Admins can tune patterns via the chat.tools.edits.autoApprove setting. Terminal auto-approve controls and improved prompts continue to add guardrails around agent-run commands.

Productivity: AGENTS.md, Prompt Improvements, Changed Files UX

  • AGENTS.md: When present at the workspace root, this file is automatically used as context for chat requests. The feature is enabled by default and can be toggled with chat.useAgentsMdFile. See agents.md for background.
  • Prompt files: You can reference custom chat modes from prompt files and configure which prompts appear in the Chat welcome view.
  • Changed files list: Now collapses by default, removes accepted items automatically, and shows counts and line deltas for a cleaner flow while working with agent suggestions.

MCP Updates

  • VS Code now reads MCP server instructions and includes them in its base prompt.
  • MCP auto-discovery is disabled by default (re-enable via chat.mcp.discovery.enabled).
  • The previous chat.mcp.enabled toggle has been migrated to chat.mcp.access with clearer options (all or none).

Other Notables

  • Accessibility: New options for editor tab index display and tab bar scrollbar visibility.
  • Source Control: Preview and migrate Git worktree changes back to the current workspace.
  • GitHub Pull Requests extension updates, plus additional terminal window ergonomics.

For model exploration and bring-your-own model options, see the AI Toolkit documentation (Explore models in AI Toolkit).

About the Author

David Ramel is an editor and writer at Converge 360.

comments powered by Disqus

Featured

  • VS Code 1.123 Adds Agent Session Sync, 1M Context Windows

    Microsoft released Visual Studio Code 1.123 on June 3, adding agent-focused features, larger model context support, integrated browser updates and a new delay for some automatic extension updates.

  • Copilot Billing Shock Hits Developers

    Developer complaints about GitHub Copilot's new usage-based billing model have centered on unexpectedly rapid AI credit consumption, and neither GitHub nor Microsoft has responded directly to the backlash, though they have previously published guidance to lessen model usage costs.

  • Hands On with GitHub Copilot App Technical Preview: Turning a Blazor Issue into a PR

    GitHub's brand-new Copilot desktop app, in technical preview, handled a small Blazor issue from planning through pull request creation, but the hands-on test also showed why developers still need to verify agent work in the running app before merging.

  • At Build 2026, Microsoft Sets Up Windows as an OS for AI Agents

    Microsoft's Build 2026 Windows developer announcements point to a broader platform strategy for agentic AI, spanning terminal workflows, local models, app-building skills, Cloud PCs and operating system-level containment.

Subscribe on YouTube