Gemini 4 Carbon opens to all paying developers at Argon's price
Google made Gemini 4 Carbon generally available at $2 per 1M input tokens and $10 per 1M output through Nov. 30, then $4 and $20.

Google has opened Gemini 4 Carbon, which it calls its strongest model, to every paying developer at the same introductory price as Gemini 4 Argon. Google says Carbon uses more tokens than Argon on a task, and the price doubles on Dec. 1.
Google made Carbon available to developers, enterprises, scientists and consumers on Nov. 10, Koray Kavukcuoglu, Google’s chief AI architect, wrote in a Google blog post.
Carbon Moves From Limited Access to General Availability
Carbon is now open to all paying developers and enterprises through the Gemini API in Google AI Studio and through Gemini for business. Two weeks earlier, Google had released it only to a set of cybersecurity defenders through its Fairwind Program, a limited-access release.
The API model ID is gemini-4-carbon, and the model is not on the Gemini API free tier. Google is also rolling Carbon out to all Google AI Ultra subscribers in Google Antigravity, the Gemini app and AI Mode in Search.
Carbon Costs the Argon Rate Until Nov. 30
Google’s Gemini API pricing page, read Nov. 11, lists Carbon at $2.00 per 1M input tokens and $10.00 per 1M output tokens through Nov. 30. From Dec. 1, 2026, the rates are $4.00 and $20.00. Every token the model processes is billed, including thinking tokens, and output pricing includes them.
Cached input costs $0.10 per 1M tokens through Nov. 30 and $0.20 after. Carbon supports implicit caching only. Explicit caching is not supported.
Google Points Carbon at the Hardest Tasks
Google says Argon “remains the ideal choice for the majority of software engineering and knowledge work.” Carbon is built for the hardest tasks, such as complex multi-step coding and debugging and open-ended scientific research, “using more tokens than Argon to explore thoroughly, reason deeply, and run extended verification loops,” the Nov. 10 post said. The pricing page describes Argon as delivering “faster responses and lower token spend per turn compared to Gemini 4 Carbon.”
Google’s Nov. 10 post says Carbon ranks first on CursorBench 4.0, Perplexity’s WANDR and Zapier’s AutomationBench “at more than 2X lower cost per task compared to the next best external model.”
Google is also previewing a /swarm feature in Google Antigravity that turns one prompt into a coordinated team of agents. Developers can assign Carbon to the roles that need the deepest reasoning and pair it with faster, lighter models.
Our View
We think the detail that sets the bill is Google’s own statement that Carbon uses more tokens than Argon. The per-token rate is the same and thinking tokens bill at the output rate. A developer who moves a job from Argon to Carbon therefore pays the same rate on a larger token count. From Dec. 1 that rate doubles.
Developers and the teams that pay API bills should send only the hardest jobs to Carbon and keep routine work on Argon, as Google recommends. Budgets should use the Dec. 1 rate, since tests run now are priced at half of what the same traffic will cost after Nov. 30.
Carbon pays for itself only if it finishes a hard task in fewer attempts than Argon would need.
