Grok 4.6 is now available in GitHub Copilot, adding xAI’s latest reasoning model to the coding service for paid individual and organizational plans. The August 14 release gives developers another option for agentic coding and complex, multi-step work, although availability is rolling out gradually.

GitHub said Grok 4.6 will be offered to customers on Copilot Pro, Pro+, Max, Business and Enterprise plans. Developers will be able to select it through the model picker in Visual Studio Code, Visual Studio, Copilot CLI, the Copilot cloud agent, the GitHub Copilot app, JetBrains IDEs, Xcode and Eclipse.

The model carries general-availability status and is billed at xAI’s provider rates through Copilot’s usage-based system. GitHub lists prices of $2 per million input tokens, $0.50 per million cached input tokens and $6 per million output tokens for prompts of up to 200,000 input tokens. Rates double for the long-context tier above that threshold.

GitHub describes Grok 4.6 as a model for agentic coding and complex workflows. The company said its internal testing produced strong results on terminal-based tasks in Visual Studio Code and Copilot CLI, particularly during longer assignments requiring sustained reasoning and tool use. Those performance statements are GitHub’s internal findings; the announcement did not provide test data that would allow readers to independently verify them.

More Choice for Complex Coding

The addition expands Copilot’s multi-provider model lineup and lets developers choose Grok 4.6 without leaving their existing editor, command-line or GitHub workflow. For teams already using Copilot, that can reduce the need to configure a separate xAI account or coding client when testing the model against repository-level tasks.

The release may carry the most weight for developers working on long-running, tool-driven assignments, such as codebase exploration, debugging and coordinated edits across several files. Still, model quality can vary by language, repository and prompt. GitHub’s internal characterization does not establish that Grok 4.6 will outperform other Copilot models in a team’s own projects.

Grok’s Rapid Copilot Upgrade

The Grok 4.6 release follows the addition of Grok 4.5 to Copilot on July 28, 2026. GitHub had described that earlier model as suitable for agentic coding and multi-step workflows, with availability across the same broad group of Copilot plans and development environments.

The short interval between releases reflects the pace of model changes inside Copilot. GitHub now offers models from xAI and several other providers, while charging usage according to the selected model’s token rates. In this system, one GitHub AI credit is valued at $0.01, and charges above included allowances depend on the amount and type of tokens consumed.

“Grok 4.6, xAI’s latest reasoning model, is now rolling out in GitHub Copilot.”

— GitHub

Performance Evidence Remains Limited

It is not yet clear when every eligible account will receive access, because GitHub gave no completion date for the gradual rollout. The announcement also supplied no detailed benchmark methodology, latency figures or comparisons with competing Copilot models. Developers will need project-specific testing to judge code quality, speed and cost, while organizations must decide whether xAI model access fits their internal policies.

Rollout and Team Testing Continue

Eligible users should watch the Copilot model picker as access expands. Business and Enterprise administrators can review the model policy before activating it, while development teams can compare Grok 4.6 against existing models on representative repositories and monitor token spending. GitHub has not announced a date for completing the rollout.

Source: xAI

Key Questions

When did Grok 4.6 become available in GitHub Copilot?

GitHub announced the rollout on August 14, 2026. The model is generally available, but the deployment remains gradual, so it may not appear for every eligible user at the same time.

Which Copilot plans include Grok 4.6?

GitHub lists Copilot Pro, Pro+, Max, Business and Enterprise as eligible plans. Access depends on the rollout reaching the account and, for managed organizations, on administrator approval.

Where can developers use the model?

Grok 4.6 is being added to the model picker in Visual Studio Code, Visual Studio and Copilot CLI, along with Copilot’s cloud agent and app. GitHub also lists JetBrains, Xcode and Eclipse.

Does an organization need to activate Grok 4.6?

Yes. For Copilot Business and Enterprise, administrators must activate the Grok 4.6 policy in Copilot settings. GitHub says the policy is disabled by default.

How is Grok 4.6 usage charged?

Copilot bills Grok 4.6 at provider token rates under usage-based billing. The standard tier costs $2 per million input tokens, $0.50 per million cached input tokens and $6 per million output tokens; the long-context tier costs more.

Source: xAI