Technical Trigger
The introduction of model selection for Claude and Codex agents on GitHub Copilot is facilitated through the update of the Copilot cloud agent, allowing users to select a model when kicking off a task. The specific models available for Claude are Anthropic models, including Claude Sonnet 4.6, Claude Opus 4.6, Claude Sonnet 4.5, and Claude Opus 4.5, while Codex utilizes OpenAI models, including GPT-5.2-Codex, GPT-5.3-Codex, and GPT-5.4.
Developer / Implementation Hook
Developers can take advantage of this update by enabling access to third-party coding agents through their existing Copilot subscription. For Copilot Business or Enterprise subscribers, the relevant policy for Anthropic Claude or OpenAI Codex must be enabled by the administrator. Additionally, the user or organization that owns the repository must enable the agent from Settings > Copilot > Cloud agent.
The Structural Shift
The introduction of model selection for Claude and Codex agents represents a shift towards more flexible and adaptable coding assistance, allowing developers to choose the best model for their specific needs.
Early Warning — Act Before Mainstream
To act on this change, developers can: 1. Review the available models for Claude and Codex and determine which ones best suit their coding tasks. 2. Enable access to third-party coding agents through their Copilot subscription and settings. 3. Explore the documentation on third-party agents to understand how to effectively utilize the new model selection feature.