Rumored DeepSeek Cyber Model Benchmarks Surface Online

Rumors and leaked benchmark screenshots suggest DeepSeek has an internal cybersecurity model scoring up to 96.8 on ExploitBench, though analysts say it may be a repackaged guardrail-free version; third-party tests show its 1017 model reaches 95.36% pass@3 on CyberGym.

2026-10-30 ~ 2026-10-31 · 4 related posts

• Rumored DeepSeek Flash-Cyber benchmarks leak, claiming ExploitBench 96.8 — teortaxesTex · 2026-10-30
• Leaked 'DeepSeek-cyber' tease: open DeepSeek models near frontier on cybersecurity CVE benchmarks — teortaxesTex · 2026-10-30
• Leaked DeepSeek Cyber Scores Analyzed: Possibly Guardrail-Free 1017, Not a New Model — teortaxesTex · 2026-10-30
• DeepSeek 1017 gains just 0.6% on CyberGym as Flash tops leaderboard at $0.23 per task — teortaxesTex · 2026-10-31