Learn about this first
Google released and globally rolled out its frontier model Gemini 4 Argon on October 14. Google says the model is designed for complex tasks in programming, finance, law, cybersecurity, and science. It can directly generate shareable web applications, printable 3D files, slides, and reports from task decomposition, and is said to offer significant improvements in spatial reasoning, mathematical precision, and visual programming.
Availability varies by user type: Google AI Ultra and Pro subscribers can use it in the Gemini app and Search AI Mode; Ultra subscribers can also use it in Antigravity; developers and enterprises must access it through the paid tiers of the Gemini API and Gemini platform. Free users and Plus subscribers are not included in the rollout and can still use only the Flash series. The introductory API price is $2 per million input tokens and $10 per million output tokens, rising to $4 and $20 after November 30, 2026. Scientists and researchers can apply for six months of free access to Antigravity’s highest usage tier, while cybersecurity defenders can apply to join the Fairwind Program. Both require eligibility review and screening.
Google says the model solved all 12 end-to-end simulated cyberattack scenarios, compared with 5 for the previous generation. Sundar Pichai said token usage approached 25 trillion within 72 hours of launch, about 3 times the amount for Gemini 3.0 over the same period, and said capacity constraints were encountered while serving the model. Early tester indigo believes Argon has helped Google catch up with the top tier again but has not surpassed its rivals, while Antigravity remains six months to a year behind.
AI-generated from reports · updated 14 days agoLatest developments
Gemini 4 Argon processed approximately 25 trillion tokens within 72 hours of launch
Event progress14
Gemini 4 Argon processed 25 trillion tokens within 72 hours of launch
Rohan Paul: Sundar Pichai says Google Argon reached 25 trillion tokens in usage within 72 hours of launch, 3 times the amount for Gemini 3.0 over the same periodBrin says AI is moving toward self-improvement
IT Home · Artificial Intelligence: Google’s Brin says Gemini is moving toward the recursive self-improvement (RSI) stageGemini 4 Argon becomes available globally to Pro/Ultra users
Gemini: Gemini 4 Argon is available worldwide to Google AI Pro and Ultra subscribersSergey Brin and others discuss Gemini 4 Argon
Logan Kilpatrick: Sergey Brin and others discuss Gemini 4 ArgonGemini 4 Argon released with a unified million-token context window
AI Watts by Karl: Hands-on test of Gemini 4 Argon’s writing ability: How does it perform after unifying the million-token context window?Hands-On Review: Gemini 4 Argon Brings Google Back to the Top Tier, but Antigravity Still Lags
indigo: Hands-on evaluation finds that Gemini 4 Argon has helped Google catch up with the top tier again, but Antigravity remains behindOpenCode launches the Gemini 4 Argon model
opencode: Gemini 4 Argon launches on OpenCode GoGoogle Antigravity Multi-Agent Generation of Complex Architecture
Sundar Pichai: Google Antigravity’s multi-agent system generates complex architectureGemini’s macOS app integrates the Antigravity coding interface
Gemini: Gemini 4 Argon brings Google Antigravity’s agentic coding interface to the Gemini app for macOSGemini 4 Argon supports designing hardware from sketches in Antigravity
Google AI: Design hardware from sketches with Gemini 4 ArgonGoogle DeepMind announces ways to access Gemini across multiple scenarios
Google: Google opens applications for the highest Antigravity tier for freeManus launches the Gemini 4 Argon model
Manus: Gemini 4 Argon launches on ManusGemini 4 Argon rolls out to Ultra subscribers
Ars Technica: AI: Gemini 4 Argon starts rolling out to paying users on the Gemini app and AntigravityGemini 4 Argon is now officially available
Google Blog: AI: Google releases Gemini 4 Argon, a frontier model for complex engineering and research tasks