AI · 5h ago
OpenAI released GPT-6.1 Terra to users on four ChatGPT plans and through its API, saying it meets the company’s Critical cybersecurity threshold. The desktop-operating model is available in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu users; API prices are 50% below GPT-5.6 Terra’s.
In OpenAI’s framework, Critical means that, when connected to tools, a model can find and build working exploits for unknown flaws in hardened critical systems without human involvement, or plan and carry out a new attack from a high-level goal. This is a capability assessment, not evidence Terra has carried out attacks. OpenAI says it retains protections developed for GPT-6 Astra, but the source provides no independent validation of how those controls perform in deployment.
For teams letting Terra operate desktop software or other tools, the open question is whether their own permissions and safeguards constrain those capabilities; the release reports no deployment incident or specific control failure.
2 sources covering this story
OpenAI's GPT-6.1 Terra is safer overall, but shows more destructive-action flags in tests
OpenAI confirmed GPT-6.1 Terra is safer overall than some of its older models, but its safety testing also found other problems.
OpenAI released GPT-6.1 Terra, a model that operates desktop software, to all Plus, Pro, Enterprise, and Edu users in ChatGPT Work and Codex, and to
Part of the PlainSec briefing for 2026-10-27