Back to news
launchGoogle2026-11-11

Google opens Gemini 4 Carbon to all paying developers, with introductory pricing of 2 dollars per million input tokens and 10 dollars per million output tokens

Google opened Gemini 4 Carbon to all paying developers and enterprise customers, with an introductory price of 2 dollars per million input tokens and 10 dollars per million output tokens, rising to 4 and 20 dollars after November 30.

Google announced on November 10 that its frontier model Gemini 4 Carbon is now open to developers, enterprise customers, researchers and consumers. The model was first announced on October 28 and had been limited to a small number of cybersecurity defenders through the Fairwind program for two weeks. Koray Kavukcuoglu, senior vice president and chief AI architect at DeepMind, said Carbon is the strongest reasoning model Google has built, aimed at complex software engineering, multi-step agent tasks, scientific research and cybersecurity. Compared with Gemini 4 Argon released at the same time, Carbon spends more inference tokens on complex tasks for hypothesis exploration, result verification and edge case checking. It is available through the Gemini API and Google AI Studio to all paying developers and to enterprise customers. The model ID is gemini-4-carbon, accepting text, image, video, audio and PDF input with text output. The unified context window is 1048576 tokens, standard synchronous requests allow up to 65536 output tokens, and the Long Decode mechanism extends maximum output to 1048576 tokens. Low, medium and high reasoning effort levels are supported. Pricing matches the introductory price of Argon at 2 dollars per million input tokens and 10 dollars per million output tokens, running through November 30, 2026, after which the price becomes 4 and 20 dollars. Google says Carbon ranks first on CursorBench 4.0, Perplexity WANDR and Zapier AutomationBench, with per-task cost less than half that of the next best external model.

Gemini 4 CarbonGoogleGemini模型定价上下文窗口智能体