Pular para o conteúdo
← Back to Skalablog

Published article

How to test Gemini 4 Argon in Antigravity

Software EngineeringGeminiChatGPTClaude

One number separates Gemini 4 Argon from its rivals: a 1 million token output limit. Google calls the model its most capable for complex tasks, and third-party evaluation placed it near the top of the field rather than clearly above it. Subscribers on Google AI Pro and Ultra get it in the Gemini app and Search AI Mode; Antigravity coding needs Ultra.

FAQ

  • Is Gemini 4 Argon available for free? No. Access starts on the Google AI Pro and Ultra paid plans in the Gemini app and Search AI Mode, and API usage is billed per token. Ultra is required for Antigravity coding access.
  • What is special about the 1 million token output limit? The ceiling applies to generated output rather than combined input and output, which Google presents as a first for frontier models. A full-length generation costs $10 at the introductory output rate and $20 after the price increase on 2026-12-01.
  • Does Gemini 4 Argon beat Claude ChatGPT at coding? Google reports state-of-the-art coding benchmark results, and independent aggregation placed the model near the top rather than clearly first. Hands-on tests found working game mechanics but below-competitor visual design.
  • When does the API price increase? Google set introductory rates of $2 per million input tokens and $10 per million output tokens, rising to $4 and $20 from December 2026. Budget any long-output workload against the later rate.

Source video