Vulnerability Exploitability: GPT-5.5 vs Claude Benchmarks

This report evaluates GPT-5.5, Claude Opus 4.7, and Claude Sonnet 4.6 on real-world vulnerability validation workflows (CVE test cases and live web-app reports), finding narrow accuracy differences but meaningful differences in speed, determinism, and reasoning efficiency; common blind spots involve judging patch completeness and separating fabricated or overstated reports from true impact, and investing in tooling/prompt engineering improved benchmark accuracy to ~98%.