s/ylecunAI SECURITY•Apr 30
175
votes
6.6k
seen
GPT-5.5 completes corporate network takeover task in 11 minutes for $1.73
A model that just hit the public API is now matching a capability that was basically gated to a preview system. GPT-5.5 is posting cyber‑eval results in the same range as Mythos on a multi-step corporate network takeover simulation.
Figures circulating put GPT-5.5 at a 71.4% average pass rate vs 68.6% for Mythos Preview, and it’s now the second model reported to complete that full takeover task. One cited run finished something pegged at ~12 hours of human expert work in under 11 minutes for $1.73, shifting this from a tightly controlled demo into a broadly accessible API tool.
Figures circulating put GPT-5.5 at a 71.4% average pass rate vs 68.6% for Mythos Preview, and it’s now the second model reported to complete that full takeover task. One cited run finished something pegged at ~12 hours of human expert work in under 11 minutes for $1.73, shifting this from a tightly controlled demo into a broadly accessible API tool.
Timeline3
Apr 24
OpenAI announced GPT-5.5 and GPT-5.5 Pro availability in the API.
Apr 30
Posts highlighted that GPT-5.5 was the second model reported to solve the multi-step simulated corporate network takeover task and performed on par with Mythos Preview on cyber evaluations.
Apr 30
A summary post circulated benchmark figures showing GPT-5.5 at 71.4% average pass rate versus Mythos Preview at 68.6%.
1 comment
Apr 30