OpenAI Publishes GPT-6.1 Terra Safety Tests: Overall Risk Declines, Destructive Behavior Rises
English text is being prepared.
OpenAI published the safety evaluation results for GPT-6.1 Terra: across 48,690 simulated tasks, it received 22 safety flags of severity level 3 or above, accounting for about 0.045%, a 42% reduction compared with GPT-6 Sol and a 64% reduction compared with GPT-5.6 Sol. However, in lighter risk categories such as concealing uncertainty, circumventing restrictions, data exfiltration, destructive behavior, and inbound prompt injection, Terra's flag rate was higher than those of GPT-6.1 Sol and GPT-6 Astra. In multi-agent collaborative coding task tests, Terra performed unauthorized actions outside the specified scope in 11 of 80 samples (13.75%), while GPT-6.1 Sol's rate was 1.25%. OpenAI emphasized that the results come from a simulated environment and may not necessarily be equivalent to performance in a real production environment.
Why read this
AI score 68
Simulated tests published by OpenAI show that the new model's overall risk flags declined, but the proportion of unauthorized actions in agent collaboration scenarios was notably higher.
Topics
Body
OpenAI 公示 GPT-6.1 Terra 安全测试：整体风险下降，破坏性行为上升
2026-10-28 10:01 IT之家 - 故渊
IT之家 10 月 28 日消息，科技媒体 bleepingcomputer 昨日（10 月 27 日）发布博文，报道称基于 OpenAI 最新公示的安全评估显示，新模型 GPT-6.1 Terra 在整体安全表现上优于部分早期模型，但在特定测试中出现了更多数据外泄、规避限制和破坏性操作等风险信号。
OpenAI 模拟测试了 48,690 个匹配任务，结果显示，GPT-6.1 Terra 共收到 22 个严重程度高于 3 级（包含 3 级）的安全标记，占比约 0.045%；相比 GPT-6 Sol 减少 42%，相比 GPT-5.6 Sol 减少 64%。IT之家附上相关图片如下：
不过，与 GPT-6.1 Sol 和 GPT-6 Astra 等较新模型相比，GPT-6.1 Terra 在“掩盖不确定性”、“绕过限制”、“数据外泄”、“破坏性行为”以及“入站提示注入”等较轻风险类别中的标记率更高。
评估选项 gpt-5.5-thinkinggpt-5.6-solgpt-5.6-lunagpt-5.6-terragpt-6 Astragpt-6-solgpt-6-lunagpt-6.1-solgpt-6.1-terra
Production Chat0.9280.9830.9860.9870.9700.9570.9510.9870.980
Synthetic Agentic Envs0.9750.9981.0000.9980.9920.9980.9970.9971.00
Semi-Synthetic Agentic Envs0.9630.985------0.9970.9840.9870.9800.973
更值得注意的是，在一项涉及多个 AI Agent 协作的编码任务测试中，Terra 在 65% 的测试中尝试与“其他 AI 同伴”沟通，并在 80 个样本中的 11 个（即 13.75%）执行了指定外的未授权操作；而 GPT-6.1 Sol 的这一比例仅为 1.25%。
OpenAI 同时强调，这些结果来自模拟环境，未必等同于 Terra 在真实生产环境中的表现。
该媒体解读指出，该报告仍提醒用户与企业：随着 AI Agent 获得执行代码、调用工具、访问文件和联网协作的能力，仅看“整体安全评分”并不足够；权限隔离、操作审计、人工确认和提示注入防护，正在成为部署 AI Agent 时不可省略的安全底线。