Service eligibility and regional restrictions
PuppyIP serves only compliant overseas businesses and their authorized personnel. Proxy services are not available in mainland China. The service may only be used for lawful business activities outside mainland China. Use of this service within mainland China is prohibited.
Hosting a proxy IP or server overseas does not change these restrictions. The service must not be provided to end users in mainland China through relaying, forwarding, sharing or resale. Before use, read the Terms of Service.
Key Takeaways
- The October 13 announcement names Copilot Pro+, Max, Business and Enterprise, across eight clients or surfaces. General availability does not mean every eligible account already has access.
- Business and Enterprise administrators control model access. A disabled default policy or an explicit restriction on Argon can affect whether members see it.
- Current usage-based promotional rates are $2 per million input tokens, $0.10 for cached input and $10 for output, until December 1, 2026. The exact cutoff timezone and later rates are not established in this table.
- One GitHub AI credit equals $0.01, with included allowances and extra charges depending on the plan. Users remaining on legacy annual request-based Pro or Pro+ plans do not receive new models or features; an old multiplier does not unlock Argon.
Check the Copilot plan attached to your account
On October 13, 2026, GitHub announced general availability, or GA, of Gemini 4 Argon in Copilot. The announcement names Copilot Pro+, Max, Business and Enterprise, rather than Copilot Pro, Free or Student. Users remaining on legacy annual request-based Pro or Pro+ plans do not receive new models or features.
This release covers coding, agentic and reasoning tasks inside Copilot. Google’s Gemini app and API are separate access paths. A Gemini subscription, Google API balance and GitHub Copilot license or allowance should not be treated as a single entitlement.
Rollout is gradual, so an eligible account may not see the model yet. These sources do not establish a complete country list, a completion time for every account or guaranteed individual access. Check your license and client rather than treating the GA label as confirmation that your account has received it.
Look for Argon in the client you actually use
The announcement lists Visual Studio Code, Visual Studio, Copilot CLI, GitHub Copilot coding agent, GitHub Copilot app, JetBrains IDEs, Xcode and Eclipse. Open the model picker in your chosen surface and look for Gemini 4 Argon. If it appears and you select it, check the model shown for that session.
Keep your IDE and Copilot extension or plugin updated. The current model support table lists Visual Studio 17.14.22 or 18.1.0 for Argon. Its minimum versions for VS Code, JetBrains, Xcode and Eclipse remain TBD, meaning not yet determined, rather than every older version being supported.
Capabilities also vary by surface and model. This announcement does not promise that selecting Argon enables Auto model selection, a one-million-token context window or configurable reasoning levels. Extended capabilities documented for other models should not be carried over to Argon.
For an organization account, check the model policy
Copilot Business and Enterprise administrators can manage Argon through model policies in Copilot settings. New models are enabled under the default model availability policy unless an administrator has turned off the global default or explicitly disabled this model. Members should not assume that a public launch overrides those settings.
Delegate to Default Policy means a model still inherits the default, as distinct from an explicit allow or disable setting. Give your administrator the relevant plan, client and model-picker observation to check. Do not change enterprise governance or network settings as a substitute for confirming model access.
The support documentation also explains that enterprise data residency or FedRAMP restrictions can limit models that do not satisfy those policies. These sources do not independently establish that Argon meets your enterprise’s requirements. An administrator should check the actual model and data-handling terms; changing location or connection does not replace permission.
Read the token rates and the promotion deadline together
Tokens are the units used to measure content processed by a model. Input tokens are sent to the model, output tokens are generated by it, and cached input is context reused by the model. An interaction is charged using the applicable categories, so counting the words in your question alone does not establish its cost.
The current Copilot pricing table lists promotional Argon rates of $2 per million input tokens, $0.10 per million cached input tokens and $10 per million output tokens. Its footnote says the promotion lasts until December 1, 2026. Treat this as a temporary rate: the page does not establish the cutoff timezone or list Argon’s rates after the promotion.
Copilot converts token costs into GitHub AI credits, with one credit equal to $0.01 USD. Individual plans have different included allowances. Business and Enterprise per-user allowances are pooled at the billing entity level. Additional usage beyond the included allowance is charged according to the applicable billing rules.
Estimate one task, then check your billing model
As a hypothetical, unexecuted example, suppose an Argon task consumes a total of 100,000 ordinary input tokens and 20,000 output tokens, with no cached input. At the current promotional rates, its model usage value is $0.20 plus $0.20, or $0.40, equivalent to 40 AI credits. No request was made, and this is not a statement of your additional amount due.
A real interaction may resend context, generate a different amount of output or use cached input. Check your usage record for the actual token categories and counts. Remaining included credits can cover usage; if the allowance is exhausted, your budget and additional-usage policy determine whether you can continue and what is charged.
Legacy annual plans have an additional access restriction. GitHub moved to usage-based billing on June 1, 2026, but some existing annual Pro and Pro+ subscribers remained on request-based billing. The current legacy document explicitly says these users do not receive new models or features, and its multiplier table does not list Argon. Check whether your account is still on that path; the Pro+ name or an invented multiplier is not evidence of access.
Confirm access and budget before starting real work
Check the license, client version, administrator model policy and model picker, followed by the billing model, remaining allowance and budget. In an unexecuted example, you might ask Argon to inspect a change across several files: define a clear scope and acceptable usage first, then decide whether to start.
GitHub describes strong results on complex coding tasks in early testing and says Argon checks edge cases. Those are publisher claims. They do not replace review of generated code and test results, or establish that a particular repository will be faster, cheaper or correct on the first attempt.
If you subscribe, or have subscribed, to Copilot through GitHub Mobile on iOS or Android, the current pricing documentation says purchasing additional AI credits is unavailable. Check the options offered by your account if usage runs out. Model availability does not imply unlimited use. No personal bill, enterprise configuration or real model interaction was inspected.
Sources
- GitHub announcement: Gemini 4 Argon in GitHub Copilot, October 13, 2026
- GitHub Docs: model token pricing, AI credits, promotion and legacy annual billing
- GitHub Docs: supported models, client versions, plans and default enablement
- GitHub Docs: legacy annual request billing, model multipliers and restrictions on new models
Frequently Asked Questions
Does a Gemini Pro or Ultra subscription unlock Argon in Copilot?
The announcement specifies GitHub Copilot plans: Pro+, Max, Business and Enterprise. A Google Gemini subscription or API balance is not evidence of that Copilot entitlement. Check the accounts and usage rules for each product separately.
My plan qualifies, but Argon is missing. Should I reinstall?
First check whether you remain on legacy annual request-based billing: GitHub explicitly says those users do not receive new models or features. Other eligible accounts should check gradual rollout, client updates and organization policy. Visual Studio has a listed minimum version, while several other IDE requirements remain TBD. Reinstalling is not an established fix for every missing model.
Does the $2 input rate mean an immediate charge on top of my subscription?
Not necessarily. Current usage-based model costs are converted into AI credits and settled against included or pooled allowances, budgets and additional-usage rules. Legacy annual request-based users do not receive new models or features. The hypothetical amount is not your individual additional bill.
What will Argon cost after December 1?
The current table establishes promotional rates until December 1, 2026, but not the cutoff timezone or subsequent Argon rates. Recheck Copilot’s official pricing when making a longer-term budget. Do not substitute prices or dates from a different Gemini access path.