Posted Jul 15, 2026 at 7:04 PM UTCEExternal LinkOpenAI trained an AI for red-teaming AI.The new GPT-Red model “can break nearly all models it is pitted against,” according to an OpenAI blog post on Wednesday. OpenAI says it used GPT-Red to find vulnerabilities in GPT-5.6 Sol, a process that made it the company’s “most robust model to prompt injections to date.”Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.Emma Roth
OpenAI trained an AI for red-teaming AI.
The new GPT-Red model “can break nearly all models it is pitted against,” according to an OpenAI blog post on Wednesday. OpenAI says it used GPT-Red to find vulnerabilities in GPT-5.6 Sol, a process that made it the company’s “most robust model to prompt injections to date.” [Link: GPT‑Red: Unlocking Self-Improvement for Robustness | https://openai.com/index/unlocking-self-improvement-gpt-red/ | OpenAI]
83 words~1 min read
