Watching GPT-5.6 Sol Ultra Write a Chrome Exploit: Exploit Development as We Know It Is Dead

The author had Claude Opus build a full V8 exploit chain against Discord's bundled Chrome 138 — after a week, 2.3 billion tokens, and $2,283 in API costs, it popped calc.

The authors reverse-engineered Chrome's January 2026 fix for CVE-2026-0899, an out-of-bounds access in V8, built a trigger PoC, and defeated a stubborn CHECK to reach arbitrary read/write.

Benchmarking GLM, DeepSeek, Kimi, and MiniMax against GPT, Claude, Gemini, and Opus on 15 real Fider vulnerabilities: frontier models still lead, but open-weight models are closing the gap fast.

Benchmarking Hacktron's scanning pipeline shows that for most applications, smaller models run repeatedly can outperform larger frontier models on cost-to-recall.

Trail of Bits details Electron CVE-2025-55305: overwriting V8 heap snapshot files bypasses code integrity checks, enabling local backdoors in Signal, 1Password, Slack, and other Chromium-based apps.

XBOW tests Grok 4.5 in its autonomous offensive-security platform: the model often hits safety guardrails but accepts correction and replans, making it a controllable component, not safe alone.