Official Source· Google DeepMind News· AI· Published Oct 28, 2026

Google introduces Gemini 4 Carbon for long-horizon engineering and cyber work

Google announced Gemini 4 Carbon on October 28, 2026, describing it as its strongest model and highlighting cybersecurity and long-horizon capabilities. The announcement includes benchmark results and API pricing; broader access for developers, businesses, and consumers is planned for later.

Google introduces Gemini 4 Carbon for long-horizon engineering and cyber work

Full Report

Google announced Gemini 4 Carbon on October 28, 2026, calling it the company’s strongest model and emphasizing cybersecurity and long-horizon work. The release extends Google’s Gemini 4 lineup. DeepMind’s model card describes the earlier Gemini 4 Argon as intended for complex workflows including coding, enterprise knowledge work, and cybersecurity defense.

Google reports that Carbon ranked No. 1 on the updated v5.0 Artificial Analysis Intelligence Index, scoring 59. It also reports scores of 88.2% on SEC-bench Pro, 54.8% on ExploitGym, and 71.0% on CWE-bench. On internal benchmarks, Google says Carbon had more than twice Argon’s success rate at generating secure code. The company says Carbon can work for days on complex, multi-step engineering tasks, including finding root causes and patching subtle bugs.

Google plans to make Carbon available to developers, businesses, and consumers through its product surfaces at a later time. Introductory API pricing is $2 per million input tokens and $10 per million output tokens, with a 95% discount for cached input tokens. After the introductory period, input and output prices are set to become $4 and $20 per million tokens, respectively.

Google says Carbon uses the same multi-layered safety and security stack as Argon. In its reported Terminal-Bench 4.0 results, successful cheating was 0.0% and attempts to cheat fell to 0.4%; on the Gray Swan Indirect Prompt Injection benchmark, Google reports a 0.0% attack success rate. These are company-reported evaluation results, not guarantees for every use case.

Why This Matters

Carbon pairs Google’s stated engineering and cybersecurity focus with published token prices and safety-evaluation results, giving prospective users concrete points to weigh when assessing capability and cost. Wider access will depend on Google’s later rollout.

Evidence

Start with the primary evidence, then review supporting sources and context.

1 items

Primary Evidence

1
  • OfficialGoogle DeepMind NewsDocumentOriginal

    Gemini 4 Carbon: Our next frontier model

Recommended signals

Recommended by brand, topic, and recent related coverage

No more recommended signals right now

Official statement

Open original source