SEC-Bench Pro
2 models tested · Updated 2026-07-09 · Verified sources only
GPT-5.6 Sol Ultra leads at 74.3%
1
OpenAI · Blog/OpenAI · 2026-07-09
Ultra mode (4 parallel agents); beats GPT-5.6 Sol (71.2%) and GPT-5.5 (45.8%).
74.3%
2
OpenAI · Blog/OpenAI · 2026-07-09
SEC-Bench Pro PoC generation; 1.6x GPT-5.5 (45.8%) at improved latency.
71.2%