GitHub Copilot: Grok 4.7 Now Available
GitHub CopilotGitHub Copilot added Grok 4.7, xAI's newest reasoning model, to its model picker across eight surfaces including Visual Studio Code, Visual Studio, Copilot CLI, the GitHub Copilot cloud agent, the GitHub Copilot app, JetBrains, Xcode, and Eclipse. Built on Grok 4.6, the model is tuned for agentic coding and complex, multistep workflows, and is rolling out gradually to Copilot Pro, Pro+, Max, Business, and Enterprise subscribers. The model is billed under usage-based billing at provider list pricing, and Business and Enterprise administrators can allow or block it through Copilot's model policy settings.
Key Takeaways
- Grok 4.7 targets agentic workflows specifically, built on Grok 4.6 with training weighted toward harder, longer-running tasks and more careful self-verification rather than quick single-turn answers.
- The model launched across eight developer surfaces at once, including VS Code, Visual Studio, Copilot CLI, the cloud agent, the Copilot app, JetBrains, Xcode, and Eclipse, rather than a single-IDE staged rollout.
- Access is limited to paid tiers, Copilot Pro, Pro+, Max, Business, and Enterprise, with the rollout arriving gradually rather than to all eligible users simultaneously.
- Enterprise and Business admins retain control, since Grok 4.7 can be enabled or blocked organization-wide through Copilot's model policy settings, just like other recent model additions.
- Pricing follows Copilot's standard usage-based billing at provider list rates, meaning teams already budgeting for premium models absorb no new billing model, only usage-based charges.
- Independent reporting notes Grok 4.7 ships with a substantially larger base model and longer context window than Grok 4.6, at unchanged per-token pricing, which is the likely driver behind Copilot's decision to add it so soon after 4.6.
Sources & Mentions
4 external resources covering this update
Grok 4.7 Joins the Copilot Model Lineup
GitHub Copilot has added Grok 4.7, xAI's latest reasoning model, to its growing roster of selectable models. The model builds directly on Grok 4.6 and is designed specifically for agentic coding and complex, multistep workflows, the kind of long-running tasks where an AI agent needs to plan, execute, and self-correct across many steps rather than just answer a single prompt.
xAI describes Grok 4.7 as trained on a harder mix of problems weighted toward work that can take many hours to complete, with an emphasis on more careful self-verification and better handling of long context. For developers using Copilot's agent modes, that combination of qualities, agentic focus and long-horizon persistence, is the primary reason the model earns a place in Copilot's picker alongside the other frontier options.
Where It's Available
Grok 4.7 can be selected in the model picker across a wide set of Copilot surfaces: Visual Studio Code, Visual Studio, Copilot CLI, the GitHub Copilot cloud agent, the GitHub Copilot app, JetBrains IDEs, Xcode, and Eclipse. That breadth mirrors how GitHub has rolled out other recent model additions, aiming to make a new model available everywhere Copilot itself is used rather than starting with a single IDE and expanding later.
Access is gated by subscription tier: the model is available to Copilot Pro, Pro+, Max, Business, and Enterprise users. As with prior model rollouts, availability is gradual, so individual developers may not see Grok 4.7 in their model picker immediately even if their plan qualifies.
Pricing and Administration
Grok 4.7 is billed under Copilot's usage-based billing at provider list pricing, the same billing structure used for other premium models in the picker. Organizations that want tighter control over which models their developers can use can manage Grok 4.7's availability through the model policy in Copilot settings. As with previous model launches, newly added models default to enabled unless an administrator has changed the organization's global default or explicitly disabled that specific model.
What This Means for Developers
The addition gives developers another frontier-class option to reach for on demanding, multistep coding tasks, sitting alongside GitHub's other recent model additions like GPT-6 Astra and Gemini 3.8 Flash. Because Grok 4.7 is optimized for agentic work in particular, it is likely to be most useful inside Copilot's agent modes and the Copilot CLI, where tasks routinely span multiple files, tool calls, and iterations rather than a single inline suggestion.