Gemini 4 Carbon

Google's cyber- and long-horizon-focused Gemini 4 model, ranked #1 on Artificial Analysis Index v5 (59) with restricted early access.

Google · Frontier/foundation models · model

Published in Feed

Overview

Google announced Gemini 4 Carbon on October 28, 2026, about two weeks after Gemini 4 Argon, positioning it as its strongest model for cybersecurity and long-horizon tasks. Google reports 88.2% on SEC-bench Pro (vulnerability finding), 54.8% on ExploitGym and 71.0% on CWE-bench (patching), plus a score of 59 on Artificial Analysis Intelligence Index v5, ranked first. Against Argon it reports more than 2x the secure-code success rate, 0.0% attack success on the Gray Swan indirect prompt-injection test (Argon: 0.7%) and about 80% fewer high-severity misalignment alerts in internal live coding. Access starts with a subset of cyber defenders and expands to trusted partners through the Fairwind Program; Google has not given a date for developer, business or consumer availability.

This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.

Key features

  • 88.2% SEC-bench Pro, 54.8% ExploitGym, 71.0% CWE-bench (Google-reported, evaluated without guardrails)
  • Artificial Analysis Intelligence Index v5.0 score of 59, first place
  • 0.0% attack success on Gray Swan indirect prompt injection (Argon 0.7%)
  • WebDev Arena up 89 Elo to 1768, ranked #4
  • Same safety stack as Argon, including chain-of-thought monitoring, with deliberate staged rollout

Use cases

  • Vulnerability discovery, exploitation testing and patching by vetted cyber defenders
  • Long-horizon agentic coding with lower overreach and cheating rates
  • Secure code generation in enterprise engineering teams once broadly available

Pricing

Introductory $2 per million input tokens and $10 per million output tokens (same as Argon), cached input 95% off; $4 and $20 per million after the introductory period.

Pricing reflects the source record and may change. Confirm current terms with the provider.

Dates and sources

Released
28 Oct 2026
Record updated
29 Oct 2026
Record last checked
29 Oct 2026

Explore the Feed

Other entries filed under frontier/foundation models; inclusion does not imply equivalent features.

  • Muse Wonder

    Meta's most capable Muse models, with active memory and long-goal focus for agents, 1M context and $4/$20 per million tokens.

  • Rho-1

    Reka's 19B omni-reasoning model that handles text, images, video and robot actions in one network, in research preview.

  • Amazon Nova 3

    Amazon's third-generation Nova frontier model family, unveiled at an AWS keynote with a 1M-token context window and a dedicated agentic tool-use variant.

Browse the Feed