OpenAI has agreed to acquire the AI security startup PromptFoo, which specializes in enhancing the security of large language model (LLM) web applications through red-teaming and custom evaluators for input-output security. This acquisition aligns with OpenAI’s current focus on enterprise adoption and the governance of AI agents in 2026, especially in light of recent upgrades to AI models that have revealed significant vulnerabilities related to prompt injection safety.
OpenAI: OpenAI is an artificial intelligence research organization that develops advanced large language models such as the GPT series and tools for enterprise applications. In early 2026, it has prioritized practical adoption, enterprise strategies, and new model releases enhancing reasoning, coding, and agentic capabilities. OpenAI is acquiring PromptFoo to integrate specialized AI security testing into its ecosystem.
PromptFoo: PromptFoo is an open-source platform and CLI tool for red-teaming, evaluating, and securing AI applications by testing LLM prompts, outputs, and vulnerabilities. It supports enterprise-grade security features like real-time threat intelligence and model-specific attack detection. As an AI security startup, PromptFoo is being acquired by OpenAI to strengthen defenses against emerging AI threats.
`json
{
“Security Testing”: “PromptFoo enables red-teaming of LLM web apps with custom evaluators for input-output security.”,
“AI Vulnerabilities”: “Recent AI model upgrades have exposed hidden safety failures in prompt injection resistance.”
}
`
Source: FirstSquawk
