Google rolls out Gemini 4 Carbon AI model, praised by Elon Musk

Google rolls out Gemini 4 Carbon AI model, praised by Elon Musk

Google introduced its Gemini 4 Carbon AI model, designed to detect software vulnerabilities and generate secure code, and received a brief endorsement from Elon Musk on X, NDTV reported on October 29, 2026. The launch took place in Mountain View, California.

Earlier internal testing of the Gemini 4 Carbon checkpoint was carried out on Google's Jetski coding platform, according to Tech Insider. On October 9, 2026, Google unveiled the model to select cybersecurity teams, KELO-AM said, and Mashable noted that the system now leads the Artificial Analysis Intelligence Index.

参与者

Google Google engineers

地点

Mountain View, California, United States

文章

2026年十月29日
3 篇总计
Google announces Gemini 4 Carbon, its new frontier AI model
Google announces Gemini 4 Carbon, its new frontier AI model

Mashable • 10月29日 10:49

Google introduced Gemini 4 Carbon, a new AI model focused on cybersecurity and long-horizon tasks, which currently leads the Artificial Analysis Intelligence Index.

可信度 72% 操纵 27% 中立
Elon Musk Calls Google's New Gemini 4 Carbon AI Model "Impressive"
Elon Musk Calls Google's New Gemini 4 Carbon AI Model "Impressive"

NDTV • 10月29日 10:19

Google introduced Gemini 4 Carbon, an AI system aimed at detecting software vulnerabilities and generating secure code, and received a brief endorsement from Elon Musk on X.

可信度 71% 操纵 8% 中立
Google's Gemini 4 Carbon looks like a diamond in the rough
Google's Gemini 4 Carbon looks like a diamond in the rough

The Register • 10月29日 05:29

Google introduced Gemini 4 Carbon, a higher-performing AI model sold at the same introductory rates as its predecessor, and highlighted its cybersecurity strengths and benchmark results.

可信度 68% 操纵 35% 中立
2026年十月28日
4 篇总计
Google's Gemini 4 Carbon Tops Every Other Model In All But Two Of Google's Benchmarks, Scoring 88.2 Percent On SEC-bench Pro Vulnerability Discovery
Google's Gemini 4 Carbon Tops Every Other Model In All But Two Of Google's Benchmarks, Scoring 88.2 Percent On SEC-bench Pro Vulnerability Discovery

Wccftech • 10月28日 19:08

Google's new Gemini 4 Carbon model reportedly outperforms most publicly available large language models on a range of internal tests, while offering lower cost per task than Anthropic's Claude Opus 5.5.

可信度 58% 操纵 48% 中立
Google announces Gemini 4 Carbon, its latest frontier AI model that you can't use
Google announces Gemini 4 Carbon, its latest frontier AI model that you can't use

Ars Technica • 10月28日 18:47

Google announced Gemini 4 Carbon, an enhanced version of its Gemini 4 Argon model that shows improved performance in code generation, vulnerability detection and agentic tasks, with a staged release planned for select users.

可信度 55% 操纵 30% 中立
Google launches Gemini 4 Carbon AI model with cybersecurity focus
Google launches Gemini 4 Carbon AI model with cybersecurity focus

KELO-AM • 10月28日 18:17

Google introduced Gemini 4 Carbon, an AI model designed to improve vulnerability detection and secure code generation, and made it initially available to select cybersecurity teams.

可信度 72% 操纵 9% 中立
Gemini 4 Carbon: 1 Employee Quote, 0 Specs [2026]
Gemini 4 Carbon: 1 Employee Quote, 0 Specs [2026]

https://tech-insider.org/ • 10月28日 17:20

Google has not announced a new Gemini 4 Carbon model; internal testing of the checkpoint is underway on the company's Jetski coding platform, according to a Business Insider report.

可信度 55% 操纵 8% 中立