Table of Contents
Google has rolled out its latest artificial intelligence model, Gemini 4 Carbon, to the general public following a release to a select group of cybersecurity defenders two weeks prior. According to Google, Carbon is its most powerful model yet and is capable of competing with leading systems from OpenAI and Anthropic.
Advanced Capabilities and Swarm Feature
Google stated that Gemini 4 Carbon surpasses the capabilities of Argon, another frontier model launched by the company just a month earlier. While Google noted that Argon remains ideal for most knowledge work and software engineering, Carbon is engineered for tasks that challenge the limits of model intelligence, such as open-ended scientific research and complex, multi-step coding.
To demonstrate Carbon’s usefulness for long-horizon work utilizing AI agents, Google announced a preview of a swarm feature on its agentic development platform, Antigravity. This feature enables users to transform a single prompt into a coordinated team of agents.
Availability and Pricing
Carbon is launching via the Gemini API in Google AI Studio and Gemini for business for paying enterprise customers and developers. It will also be accessible to all Google AI Ultra subscribers within the Gemini app, Antigravity, and AI Mode in Search, though availability for free users and lower-tier subscribers remains uncertain.
Google has priced Carbon competitively, launching with an introductory rate of $2 per million input tokens and $10 per million output tokens. This cost is approximately a fifth of OpenAI’s GPT-6 Astra and a quarter of Anthropic’s Fable 5.2 model. Even when these introductory prices double at the end of November, Carbon will remain notably less expensive per token than its frontier competitors, although token usage can vary across different models for identical tasks.
Market Impact and Benchmarks
Benchmark data highlighted by Google indicates that Carbon ranks among the most powerful models on the market across most metrics, trailing just behind Fable 5.2. This rapid rollout, following closely after the launch of Argon, signals Google’s prominent return to the forefront of artificial intelligence research after a period where frontier development was largely dominated by OpenAI and Anthropic.
