Free Gemini users and paying Plus subscribers don't get Google's new flagship model. Under the new tier structure, they only get the weaker Flash models.
Google is rolling out Gemini 4 Argon globally, available to developers, businesses and consumers starting today. It replaces the Gemini Pro class as Google's frontier series, and the company says more Gemini 4 models will follow.
Google's new flagship AI is live worldwide. Who gets it depends on your plan. Free users and Plus subscribers only get the weaker Flash models.
Vals Index: Gemini 4 Argon 68.9%, Claude Opus 5.5 66.97%, Wonder Full 64.57%, GPT-6 Astra 63.13%. Artificial Analysis Intelligence Index: Gemini 4 Argon 53 (high), Claude Opus 5.5 54 (high with fallback), Muse Wonder Full 56 (max), GPT-6 Astra 51 (high).
Terminal-Bench 4.0: Gemini 4 Argon 57.58% (Mini-SWE-agent), Claude Opus 5.5 65.15% (Mini-SWE-agent), Wonder Full 58.08% (Mini-SWE-agent), GPT-6 Astra 59.6% (Mini-SWE-agent). Vibe Code Bench: Gemini 4 Argon 91.91% (OpenHands), Claude Opus 5.5 90.29% (OpenHands), Wonder Full 86% (OpenHands), GPT-6 Astra 89.59% (OpenHands). Argon tops the Vals Index but trails Claude Opus 5.5 overall and sits behind Meta's Muse Wonder Full.
Argon is Google's new flagship. Gemini 4 Argon replaces the Gemini Pro class as Google's frontier series. The company says more Gemini 4 models will follow.
Access depends on your plan. AI Ultra and Pro subscribers get it in the Gemini app and Search's AI mode. Free and Plus users get Flash only.
The introductory price. $2 per million input tokens and $10 per million output tokens at the introductory rate. Cached inputs get a 95 percent discount.
It tops the Vals Index, not every ranking. Argon holds the top spot on the Vals Index but trails Claude Opus 5.5 overall, and sits behind Meta's Muse Wonder Full.
Google's own results are self-reported. Google says engineers using Argon raised agent-driven code submissions by more than 35 percent, and rollbacks fell nearly a third.
Google shared self-reported results. Engineers using Argon increased agent-driven code submissions by more than 35 percent, and code rollbacks dropped by nearly a third. On safety, Google says Argon recorded a 0.7 percent successful attack rate on the Gray Swan Indirect Prompt Injection benchmark.
Google is also offering 10,000 researchers the highest usage tier of Google Antigravity free for six months.