Google Carbon Report - Privacy Safety Security Advice - Total score: leaderboard
Google DeepMind Carbon safety report, Figure 5.18. 450 public single-turn questions, nine topics/50 each; two autoraters: expert helpfulness rubric and 32 high-severity pitfalls. Raw model, no safeguards/search. Positive/negative rubric components and pitfall-free rate remain separate. Metric: Score (%). Higher is better. Source: https://storage.googleapis.com/deepmind-media/gemini/gemini_4_carbon_safety_report.pdf#page=80
Metric: Score (%). Source: storage.googleapis.com. Saturation forecast: Around June 2028. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.8 Flash | 78.41 | #36 |
| 2 | Gemini 4 Carbon | 77.15 | #2 |
| 3 | Gemini 4 Argon | 76.45 | #11 |
No result here: #1 Claude Fable 5.2, #5 Claude Opus 5.5, #6 GPT-6 Astra.
Interactive version: theaggregate.ai/benchmark?slug=google-carbon-report-privacy-safety-security-advice-total-score · How It Works · Data refreshed daily, snapshot 2026-11-12.