GitHub adds Gemini 4 Argon to Copilot with gradual rollout and usage-based billing
GitHub announced general availability of Gemini 4 Argon in GitHub Copilot on October 13. The company describes the model as intended for complex coding, agentic and reasoning tasks. General availability does not mean the model is visible in every eligible account immediately: the same announcement says rollout is gradual and asks users to check again if it is missing.
The listed eligible plans are Copilot Pro+, Max, Business and Enterprise. GitHub lists model selection in Visual Studio Code, Visual Studio, Copilot CLI, the Copilot coding agent, the Copilot app, JetBrains IDEs, Xcode and Eclipse. The release does not establish access for every Copilot plan or guarantee that every interface in an organization has already received the rollout.
GitHub says the model is billed at provider list pricing under usage-based billing. Its linked pricing documentation explains that input, output and cached tokens contribute to usage, converted into GitHub AI credits. Included credit allowances vary by plan, and additional usage beyond the relevant allowance can be billed. Usage-based access should not be described as unlimited use bundled into a seat price.
Business and Enterprise administrators can manage Gemini 4 Argon through the model policy in Copilot settings. Under default model enablement, new models become enabled unless an administrator has turned off the global default or explicitly disabled that model. This is a conditional default, not a claim that every organization has authorized access or that administrators cannot restrict it.
The announcement says early testing found more solved tasks than previous Gemini models across its coding benchmarks, with the largest gains in Copilot CLI. It also describes checking work and edge cases. These are vendor statements without standardized scores in the release; they do not prove a particular team will see the same improvement or eliminate the need to review generated code.
The changelog does not give an exact token rate, a rollout completion date or a complete regional availability list. Buyers should check the current model-pricing documentation, account visibility, credit allowance and administrator policy before budgeting or adopting the model. Announced availability and benchmark claims are different from verified access, predictable costs or correct results in a production workflow.