Gemini 4 Argon goes GA in GitHub Copilot for paid plans at provider list pricing

GitHub made Gemini 4 Argon generally available in Copilot today, putting Google's newest frontier model in the model picker for Copilot Pro+, Max, Business and Enterprise users. The launch is live worldwide: Ars Technica reported the broad rollout began the same day, and The Decoder confirmed global availability through Google's API, the Gemini app, Antigravity and AI Studio. Copilot was the big gap in that rollout, and it is now closed.

What GitHub actually shipped

Argon targets complex coding, agentic and reasoning tasks. GitHub says that in early testing the model solved more tasks than previous Gemini models on every coding benchmark, with its largest gains in Copilot CLI. The company also highlights self-checking behavior: the model tests edge cases before it calls a task finished.

Two practical caveats sit directly in the changelog. Rollout is gradual, so the model may not appear in your picker yet even on an eligible plan. And Business and Enterprise administrators control access through the model policy in Copilot settings. Under default model enablement, new models turn on automatically unless the global default is off or someone explicitly disables Argon.

The metered billing changes your math

Argon isn't covered by a premium-request allotment the way some Copilot models are. GitHub bills it at provider list pricing under usage-based billing. That price is Google's, and it is currently discounted: Ars reports $2 per million input tokens and $10 per million output tokens, with a 95 percent discount on cached tokens, and says the rate rises to $4/$20 after November 30. GitHub points to its models and pricing documentation for exact billing, so verify the pass-through there before budgeting agentic work — especially anything planned past the intro window.

Where Argon fits the frontier field

Both independent reports add context the changelog leaves out. The Decoder notes Argon still trails Anthropic's Claude Opus 5.5 in independent overall rankings, and also Meta's Muse Wonder Full, which launched yesterday. Google's own internal numbers — a 35 percent lift in agent-driven code submissions and roughly a third fewer rollbacks — are self-reported, and its consumer rollout is paywalled: the Gemini app gets Argon on AI Pro ($20/mo) and AI Ultra ($99/mo), while Antigravity and AI Studio stay Ultra-only. Copilot availability widens the developer path; it doesn't change those Google-side limits.

Worth checking before you switch

Copilot Business and Enterprise users should first confirm their admin's model policy allows Argon. Then verify what the metered spend means for your organization's budget. If you're weighing Google's model against Claude or GPT inside Copilot, compare per-task cost, not just benchmarks — our AI agent tools guide covers the current options side by side.

Argon reaching Copilot matters less as a model story than as a platform story: every Copilot developer on a paid plan can now point an agent at a repo in the same picker as Claude and GPT. The rollout speed, the November 30 price step-up, and each organization's model policy are the details to check today.