Google DeepMind announces Gemini 4 Carbon frontier model
Original title: Gemini 4 Carbon: Our next frontier model
Limited access starts with cyber defenders under Fairwind, at Argon’s $2/$10 per million token prices.
Google DeepMind announced Gemini 4 Carbon, calling it its strongest model and a follow-on to Gemini 4 Argon from two weeks earlier. DeepMind says Carbon ranks first on Artificial Analysis Intelligence Index v5.0 at 59 and leads SEC-bench Pro, ExploitGym, and CWE-bench. Access starts this week for a subset of cyber defenders, with a gradual Fairwind rollout before developer, business, and consumer availability. Introductory pricing matches Argon.
Unconfirmed: Benchmark scores, uplifts, and safety gains are DeepMind’s own figures and are not independently confirmed here.
- Carbon scores 88.2% on SEC-bench Pro, 54.8% on ExploitGym, and 71.0% on CWE-bench.
- DeepMind says secure-code success more than doubled versus Argon internally.
- Successful cheating on Terminal-Bench 4.0 is reported at 0.0%, with attempts at 0.4%.
- DeepMind says it is in the U.S. voluntary pre-release testing process.
- Cached input tokens are priced at 95% off the input token price.
- Vendors
- Google DeepMind
- Organizations
- Wiz
- Countries
- United States
- AI models
- Gemini 4 CarbonGemini 4 Argon
From the original article
4 min read at deepmind.google
Gemini 4 Carbon is our strongest model with state-of-the-art cybersecurity and long horizon capabilities. This content is generated by Google AI. Generative AI is experimental Today, we’re announcing the next frontier model: Gemini 4 Carbon, our strongest model with state-of-the-art cybersecurity and long horizon capabilities.
deepmind.google
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.