Source time Edition date
Gemini 4 Carbon is now available
Gemini 4 Carbon 现已可用
Gemini 4 Carbon's availability from Google DeepMind matters to developers and enterprises choosing a frontier model, though readers must weigh migration cost and lock-in against its capabilities before committing.
Judgment
CTOs and enterprise AI platform leads should decide whether to pilot Gemini 4 Carbon on complex multi-step coding and document-heavy workflows before budgeting a migration, because it matches Argon's $2/$10 per-million-token price but only pays off where extended reasoning and verification loops replace cheaper Argon calls.
Basis
Google DeepMind's SVP and Chief AI Architect said Gemini 4 Carbon is now available to developers, enterprises, scientists, and consumers, calling it the strongest model for cybersecurity capabilities and long-horizon agentic workflows. The mechanism is tiering: Argon remains the default for most software engineering and knowledge work, while Carbon is positioned for complex multi-step coding/debugging and open-ended scientific research, using more tokens to explore, reason deeply, and run extended verification loops. Google says Carbon ranks #1 on CursorBench 4.0, Perplexity's WANDR, and Zapier's AutomationBench at more than 2X lower cost per task than the next best external model. It is available via API in Google AI Studio and Gemini for business to paying developers and enterprises at the same introductory price as Gemini 4 Argon: $2 per million input tokens and $10 per million output tokens. A preview of /swarm in Google Antigravity turns one prompt into coordinated agents, with Carbon assignable to architectural design, coordination, and auditing while paired with faster models and adversarial verifiers.
- Input price$2 / 1M tokens
- Output price$10 / 1M tokens
- Cost/task vs next bestmore than 2X lower
- MKA avg cost vs topmore than 6X lower
Unclear The source does not state how much migration effort is required to move existing Argon workloads to Carbon, nor does it give contract, data-residency, or post-introductory pricing terms.
Next Check whether Google publishes the CursorBench 4.0, WANDR, and AutomationBench task-level cost methodology and confirms how long the Argon-matched $2/$10 introductory price lasts.