"When developers choose "Auto" in the model picker, VS Code automatically selects a model for optimal performance and avoidance of rate limits. Automatic model selection is being rolled out to all GitHub Copilot users in VS Code in the following weeks. The model picker will choose between Claude Sonnet 4, GPT-5, GPT-5 mini, and GPT-4.1 and Gemini Pro 2.5, unless access to any of these models was disabled by the organization."
"When in agent mode, the agent can autonomously edit files in a VS Code workspace. This might include accidentally or maliciously modifying or deleting important files such as configuration files, which could cause negative side-effects. But this release adds a layer of security by having the agent explicitly asking for user confirmation before editing certain files. Additionally for security, this release introduces improvements to the terminal auto approve setting to enhance both security and usability."
Visual Studio Code 1.104 previews automatic model selection in chat, letting users pick "Auto" so VS Code selects an optimal model and avoids rate limits. The model picker can choose between Claude Sonnet 4, GPT-5, GPT-5 mini, GPT-4.1 and Gemini Pro 2.5, subject to organizational access. The finalized LanguageModelChatProviders API allows extensions to contribute local or cloud-hosted language models selectable in chat. Agent security is improved by requiring explicit user confirmation before agents autonomously edit certain workspace files. Terminal auto approve setting enhancements let users enable or disable the feature, improving both security and usability. Downloads are available for Windows, Linux, and Mac.
Read at InfoWorld
Unable to calculate read time
Collection
[
|
...
]