Google DeepMind Announces Gemini 4 Carbon Frontier Model

Google DeepMind has announced Gemini 4 Carbon, a frontier model built for cybersecurity defense and long-horizon engineering workflows. Building on Gemini 4 Argon, which launched two weeks ago, Carbon is being rolled out starting this week to a subset of cyber defenders and trusted partners through the Fairwind Program before a broader release to developers, businesses, and consumers.

Gemini 4 Carbon launches at an introductory price of $2 per million input tokens and $10 per million output tokens. Cached input tokens are priced at 95% off the input token price. After the introductory period expires, pricing is $4 per 1M input tokens and $20 per 1M output tokens.

According to Google, Carbon ranks #1 on the updated v5.0 of the Artificial Analysis Intelligence Index with a score of 59, and jumped +89 elo points on WebDev Arena to rank #4 with a score of 1768.

Across external benchmarks evaluated without guardrails, Google reports that Carbon achieves 88.2% on SEC-bench Pro for finding real vulnerabilities, 54.8% on the exploitation benchmark ExploitGym, and 71.0% on Collinear's CWE-bench for patching.

According to Google, on internal benchmarks Carbon achieved more than a 2x uplift in success rate over Argon in generating secure code, while the Chrome team reported Carbon performed more than 10X better than another lab's large frontier model when generating high-quality patches.

Google states that Carbon reduced successful cheating on Terminal-Bench 4.0 to 0.0%, with attempts to cheat dropping to 0.4%. In live coding environments with Google's internal teams, the company observed that high-severity misalignment alerts dropped by almost 80%. Google also says that in its internal testing, Carbon cut unsupported claims by approximately 2x over Argon, and that the model reduced overreach by 26% and under-service by 29%. On Gray Swan's Indirect Prompt Injection Benchmark, Google reports that Carbon achieved a 0.0% attack success rate, compared to 0.7% for Argon.