Google Announces Gemini 4 Carbon With Restricted Access and Top Artificial Analysis Ranking
Understand the event
Google DeepMind announced Gemini 4 Carbon with initial access limited to selected cyber defenders, while Artificial Analysis ranked it first on its Intelligence Index v5; the model is not yet publicly available. Google said access would begin this week and gradually expand to trusted partners through its Fairwind Program as it gathers feedback and refines guardrails before broader availability to developers, businesses, and consumers. Artificial Analysis gave Carbon a score of 59, including the highest score in four of nine evaluations. In its coding-agent comparison using Antigravity CLI, Carbon scored 59, behind Claude Opus 5.5 at 62. Google reported cybersecurity scores of 88.2% on SEC-bench Pro, 54.8% on ExploitGym, and 71.0% on CWE-bench, all evaluated without guardrails. Sundar Pichai said Carbon significantly improves on Argon in long-horizon tasks and has strengthened frontier safeguards. The restricted rollout limits who can try the model despite its published benchmark results. Nathan Lambert criticized Google for announcing scores for a model he could not publicly access.
Generated by AI from reports · Updated 5 minutes ago
Read next
More changes in today’s briefingGoogle DeepMind announces Genie 4 for real-time world simulation with native audioModels in this story
Claude Opus 5.5#1 overall$4 input / $20 output per million tokens
Developments
5 developments
- 10/29 01:27 (UTC+8) · 1 reportsNathan Lambert Says Gemini Models Aren’t Publicly Available When He Wants to Try ThemNathan Lambert · X:Nathan Lambert Says Gemini Models Aren’t Publicly Available When He Wants to Try Them
- 10/29 01:23 (UTC+8) · 1 reportsNathan Lambert Says Gemini May Be Back in the Agentic EraNathan Lambert · X:Nathan Lambert Says Gemini May Be Back in the Agentic Era
- 10/29 01:23 (UTC+8) · 1 reportsArtificial Analysis Ranks Gemini 4 Carbon First on Its Intelligence Index v5Artificial Analysis · X:Gemini 4 Carbon Tops Artificial Analysis Intelligence Index v5
- 10/29 01:03 (UTC+8) · 1 reportsGemini 4 Carbon Leads Key BenchmarksDemis Hassabis · X:Demis Hassabis touts Gemini 4 Carbon’s benchmark performance and safety research
- 10/29 01:00 (UTC+8) · 6 reportsGoogle DeepMind Announces Gemini 4 CarbonGoogle DeepMind:Google DeepMind announces Gemini 4 Carbon with cybersecurity and long-horizon capabilities
Report timeline
Follow the reports to explore different sides of the event.
- Josh Woodward · XPicksGoogle introduces Gemini 4 Carbon, with cyber defenders getting initial access
Josh Woodward highlights the pace of Gemini releases and positive internal sentiment. 2 weeks ago: Gemini 4 Argon Today: Gemini 4 Carbon Quoting Sundar Pichai, he says Carbon will reach cyber defenders first, followed by wider availability. Pichai reports a score of 59 and #1 ranking on Artificial Analysis Intelligence Index v5.0, improvements in long-horizon tasks over Argon, and strengthened frontier safeguards. https://twitter.com/sundarpichai/status/2115489219804352921
- Nathan Lambert Says Gemini Models Aren’t Publicly Available When He Wants to Try Them
Gdmi whenever I want to try a Gemini model it’s not publicly available. Don’t bother announcing your scores if we can’t use it. https://t.co/jfbUbrQXNJ
- Nathan Lambert Says Gemini May Be Back in the Agentic Era
I honestly don’t know how to try Gemini in the agentic era, maybe I finally need to fix that. Pretty wild they just have been out of it since December 2025 but look maybe back. https://t.co/sjord9cPwV In a quoted post, @sundarpichai says Google is introducing Gemini 4 Carbon, its next frontier model, and that it ranks #1 on the recently updated v5.0 of @ArtificialAnlys. Nathan Lambert reacts that Gemini may be back.
- Gemini 4 Carbon Tops Artificial Analysis Intelligence Index v5
Artificial Analysis ranks Google’s Gemini 4 Carbon No. 1 on its Intelligence Index v5, with a score of 59 and the top score in 4 of 9 evaluations. In a coding-agent comparison using Antigravity CLI, it scores 59, behind Claude Opus 5.5 at 62. Gemini 4 Carbon is rolling out to selected users and is not yet publicly available; Google is offering a 50% launch discount with no confirmed end date.
- Google announces Gemini 4 Carbon with cybersecurity and long-horizon capabilities
Google announces Gemini 4 Carbon as its strongest model, with state-of-the-art cybersecurity and long-horizon capabilities. Google says Carbon improves across most benchmarks, delivering strong frontier performance in autonomous coding and cyber defense. Carbon is initially available to a subset of trusted cyber defenders in Google’s Fairwind program, with access to expand as soon as possible.
- Demis Hassabis touts Gemini 4 Carbon’s benchmark performance and safety research
Demis Hassabis says Gemini 4 Carbon is SOTA across key benchmarks and combines frontier capabilities with deep safety and security research. https://t.co/TC8Y8ziPOI
- Google DeepMind introduces Gemini 4 Carbon as its next frontier model
Gemini 4 Carbon is Google DeepMind’s next frontier model. Google DeepMind says it is state-of-the-art across cybersecurity and long-horizon capabilities.
- Logan Kilpatrick · XPicksLogan Kilpatrick introduces Gemini 4 Carbon, claiming a big jump over Argon
Introducing Gemini 4 Carbon, which Logan Kilpatrick calls a big jump over Argon, released just 2 weeks ago. Kilpatrick says Gemini 4 Carbon is SOTA on cyber and long horizon tasks, ranks #1 on Artificial Analysis V5, significantly improves usability in domains like coding, and is more aligned. https://t.co/Ikk9YedjCQ
- Sundar Pichai · XPicksGoogle introduces Gemini 4 Carbon with leading intelligence and cybersecurity evaluation scores
Sundar Pichai introduces Gemini 4 Carbon, Google’s next frontier model, following Argon’s release two weeks ago. Carbon ranks #1 on Artificial Analysis Intelligence Index v5.0 at 59 and scores 88.2% on SEC-bench Pro, 54.8% on ExploitGym, and 71.0% on CWE-bench. Pichai reports significant improvements in long-horizon tasks compared with Argon and strengthened frontier safeguards. Some trusted cyber defenders will get access starting this week, followed by more partners through the Fairwind Program ahead of wider availability soon.
- Google DeepMindPicksGoogle DeepMind announces Gemini 4 Carbon with cybersecurity and long-horizon capabilities
Google DeepMind announced Gemini 4 Carbon, reporting a score of 59 on Artificial Analysis Intelligence Index v5.0 and cybersecurity benchmark results of 88.2% on SEC-bench Pro, 54.8% on ExploitGym, and 71.0% on CWE-bench, evaluated without guardrails. Access begins this week for a subset of cyber defenders and will expand through the Fairwind Program before broader availability. Introductory pricing will match Argon at $2 per million input tokens and $10 per million output tokens, with cached input tokens discounted 95%.
Keep up with AI, one email a day
Model releases, price changes and compute developments, with original sources. Free, confirmed by email, and easy to unsubscribe. Read a past issue
Event heat trend
There is not enough continuous observation data to plot a trend yet.