Home · Technology · Nov 3 archive

Google's Gemini 4 Carbon Leads in Cybersecurity and Coding Benchmarks

Updated 9:36 AM ET · 12:44 AM ETConfirmedShare via:      

Technology Desk

In Short: Google's Gemini 4 Carbon model leads in cybersecurity and coding benchmarks, outperforming competitors.

Google's Gemini 4 Carbon model has taken the lead in cybersecurity and coding benchmarks, surpassing competitors like Anthropic's Claude and OpenAI's GPT-6 Astra. In the SEC-bench Pro, Carbon achieved a score of 88.2%, while in ExploitGym it scored 54.8%, and in CWE-bench, it reached 71.0%.

On the Artificial Analysis Intelligence Index, Carbon ranks first with a score of 59, ahead of Anthropic's Claude Opus 5.5 (58) and OpenAI's GPT-6 Astra (57). The benchmark includes 206 tasks across various domains, including finance, materials science, cybersecurity, software engineering, legal, biology, and more.

Gemini 4 Carbon will launch at the same introductory price as Argon, with $2 per million input tokens and $10 per million output tokens. Cached input tokens will be priced at 95% off the input token price. After the introductory period, the rates will rise to $4 and $20, respectively.

In internal benchmarks, Carbon shows significant gains in generating secure code and demonstrates reduced misalignment and fabrication rates in live coding environments with Google's own teams. It also outperforms another lab's large frontier model by more than 10 times in generating high-quality patches.

Building on Gemini 4 Argon, which was launched two weeks ago, Carbon illustrates the rapid progress Google is making in AI capabilities. Starting this week, Carbon will be made available to a subset of cyber defenders and will gradually phase access to additional trusted partners through the Fairwind Program.

Google is continuing to gather feedback and iterate on guardrails before making Carbon available to developers, businesses, and consumers via its product surfaces as soon as possible.

With zero external libraries and 100% pure Python, Argon wrote its own 3D ray-tracing engine with a custom video tool and audio mixer to build and document a sunlit library from scratch, producing, narrating, and assembling this video with Gemini 3.8 TTS voiceover and music by Lyria 3.5.

For everyday users, Gemini 4 Argon is rolling out inside the main Gemini app and Google Search AI Mode, but only if you pay for AI Pro at $20 a month or AI Ultra starting at $99.

However, the firm also found Argon’s raw accuracy fell to 50%, five points below Gemini 3.1 Pro Preview.

Background

What's confirmed

• It shows significant gains across coding and agentic tasks as well as across key capabilities critical for cybersecurity defense, landing a firm lead in SEC-bench Pro (88.2%) for finding vulnerabilities, ExploitGym (54.8%) for exploitation, and CWE-bench (71.0%) for patching.
• Carbon ranks #1 on the recently updated v5.0 of Artificial Analysis Intelligence Index at 59.
• Gemini 4 Carbon will launch at the same introductory price 1 as Argon at $2 per million input tokens and $10 per million output tokens, with cached input tokens priced at 95% off input token price.
• As another example, the Chrome team has reported that Carbon performed more than 10X better than another lab's large frontier model when generating high quality patches.

What's still developing

• The announcement highlights Carbon’s top-tier benchmark performance and signals a clear push toward more capable, security-focused AI systems.
• "GPT‑6.1 Terra meets the Critical threshold in cybersecurity under our Preparedness Framework, and maintains the protections developed for GPT‑6 Astra," OpenAI said.
• As The Rundown reported, the new frontier model is initially limited to select vetted cybersecurity teams.
• Koray Kavukcuoglu, senior vice president at Google DeepMind and chief AI architect at Google, said Carbon represents the company's best work to date.
• No exact score split was published for that comparison, so the practical read is that the two models are functionally even on this specific cybersecurity measure.
• Gemini 4 Argon: Built to solve your hardest problems We’re giving developers, enterprises, scientists, and consumers access to our latest frontier model.

Sources

• blog.googlelink
• Bloglink
• Dataconomylink
• The Hacker Newslink
• Therundownlink
• news.lavx.hulink
• Androidheadlineslink
• Shatteredlink
• Warpbeatlink
• Bloglink
• Cyberkendralink
• Nationaltechnologylink
• WarpBeat — background on Google's Gemini 4 Carbon Outperforms Competitors in Cybersecurity and Coding link
• WarpBeat — background on Google's Gemini 4 Carbon Scores High in Cybersecurity and Coding Benchmarks link