OpenAI's Astra achieves 100% success rate on ExploitBench vulnerability tests
JiaweiLiu_ · x · 2026-09-02
Tests show that OpenAI's new model, Astra, achieved a 100% Automated Cybersecurity Exploit (ACE) success rate against all 41 CVEs in ExploitBench. To rule out contamination, an internal port using only V8 CVEs from the past 3 months was created. Astra still demonstrated a major capability increase over GPT-4o (5.6) while using significantly fewer tokens. OpenAI previously stated that Astra has reached the 'Critical' threshold under its Preparedness Framework.
More from Models
- DeepSeek V4.1 Flash enters BuildingBench top five at 1/15 of Sonnet's cost — Lianhuiq · 2026-10-18
- MOSS SFT vs Flux 3 Klein vs SD 3.0: Same-Prompt Comparison Exposes SD3's Age — ZootAllures9111 · 2026-10-18
- Quick take: GPT-6 matches Opus 5.5 at coding, far ahead at computer use — _vztu · 2026-10-18
- Open models now above Sol-level, outpacing expectations as community awaits V4.1 Pro — yacineMTB · 2026-10-18
- Codex took 3 days, Opus 5.5 nailed it in an afternoon — banned user's video protest — dotey · 2026-10-18
- Yacine MTB: Users will remember how Anthropic treated them once intelligence proliferates — yacineMTB · 2026-10-18