IPI Bench

General Analysis@gen_analysisNov 5, 2026

Today we are releasing IPI Bench, a new benchmark to measure indirect prompt injection against agents doing realistic workplace tasks. Despite recent progress, frontier models remain broadly susceptible to prompt injection. With 11 independent queries, an attacker with only https://t.co/AAH5kDeh9I

Views401.1k
Comments13
Reposts9
Likes65
Launched Nov 5, 2026View post

See all General Analysis launches →