Google Carbon Report - SEC-bench Pro - May 2026 JS 10h: leaderboard

Google DeepMind Carbon safety report, Figure 7.3. 183 historical Chromium V8 / Firefox SpiderMonkey vulnerabilities, May2026 split; Pass@1 mean of3 attempts each,10h/attempt; internal Antigravity. PoC must crash vulnerable not patched build, or judge verifies intended flaw if both crash. Metric: Pass@1 (%). Higher is better. Source: https://storage.googleapis.com/deepmind-media/gemini/gemini_4_carbon_safety_report.pdf#page=91

Metric: Pass@1 (%). Source: storage.googleapis.com. Saturation forecast: Around February 2027. 2 models tracked.

Top models

#ModelScoreOverall rank
1Gemini 4 Carbon88.2#2
2Gemini 4 Argon83.4#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-sec-bench-pro-may-2026-js-10h · How It Works · Data refreshed daily, snapshot 2026-11-12.