Google DeepMind's Gemini 4 Carbon takes the #1 spot on Artificial Analysis, two weeks after Argon
What's the deal? On 28 October 2026, Google DeepMind announced Gemini 4 Carbon, the new frontier model from Alphabet | Google, which Google calls its strongest model yet for cybersecurity and long-horizon work. It arrives two weeks after the wider rollout of Gemini 4 Argon (announced 30 September), and Sundar Pichai framed it as part of a cadence of periodic model releases.
Benchmarks: Independent evaluator Artificial Analysis ranks Carbon (high reasoning) first on its Intelligence Index v5.0 with 59, ahead of Anthropic's Claude Opus 5.5 (58), OpenAI's GPT-6 Astra (57) and Argon (50). On its Coding Agent Index, Carbon scores 59, three points behind Claude Code with Opus 5.5. In Google's own table Carbon leads most rows, including 88.2% on SEC-bench Pro (finding vulnerabilities), 54.8% on ExploitGym and 71.0% on CWE-bench (patching), but GPT-6 Astra stays ahead on FrontierSWE 2 (65.5% vs 61.2%) and Claude Opus 5.5 on OSWorld-2.1 (87.2% vs 81.5%).
Safety claims: Google says Carbon more than doubles Argon's success rate at generating secure code, cut high-severity misalignment alerts in internal coding work by almost 80%, roughly halves unsupported claims, and resisted every attack in Gray Swan's indirect prompt-injection benchmark. Its Chrome team reports Carbon produced high-quality patches more than 10x better than another lab's large frontier model.
Who can use it? Not the public yet. Starting this week, access goes to a subset of cyber defenders and then to more trusted partners through Google's Fairwind Program, and Google is taking part in the US government's voluntary pre-release testing. Pricing matches Argon's introductory $2 per million input tokens and $10 per million output tokens, rising to $4 and $20 after the introductory period. Google has not published a technical report or parameter count, and Ars Technica notes Carbon could be a post-training upgrade of Argon rather than a new architecture.
The signal: Two frontier releases in a month put Google back on top of the main independent leaderboard, with margins of only a few points over Anthropic and OpenAI. The post drew a one-word "Impressive" from Elon Musk, whose xAI competes in the same race. The defenders-first rollout continues the gated pattern now common across the leading labs.
Read more: Google · Artificial Analysis · VentureBeat · Ars Technica · Sundar Pichai on X
Image credit: Google
