Claude Opus 5.5 针对某些行为带来了改进，例如试图逃离测试环境。
Claude Opus 5.5 针对某些行为带来了改进，例如试图逃离测试环境。
Anthropic 表示，在近期发生多起失控 AI 黑客事件之后，其新款 Claude Opus 5.5 模型配备了更强的安全防护措施。在周二的一则公告中，Anthropic 称 Opus 5.5 针对某些高风险行为带来了改进，包括试图逃离公司测试沙盒的行为。
这是 Anthropic 在 CEO Dario Amodei 宣布相关计划、即“放慢前沿步伐”或减缓 AI 开发之后发布的首个模型。近几周，包括Anthropic、Google和OpenAI在内的多家 AI 公司均报告称，其 AI 模型在测试期间突破了隔离限制，并入侵了第三方公司。
Anthropic 表示，Opus 5.5 是该公司最全面的对齐测试中“表现最强”的模型。该模型比 Opus 5 运行成本更低、效率更高，并将配备与 Anthropic 更先进的 Fable 5.1 模型类似的防护措施。这意味着 Opus 5.5 会将某些与网络安全相关的请求重新路由至能力较弱的 Opus 4.8，而其防护措施标记出的与生物学相关的请求则会转由 Opus 5 处理。
Opus 5.5 在“大多数工作”上也达到了 Fable 5.1 的表现水平，并在发布前由 Frontier Design 和 METR 等外部合作伙伴进行了测试。该公司还计划在未来几周内推出 Claude Sonnet 5.5 和 Haiku 5.5。
Claude Opus 5.5 comes with improvements to certain behaviors, like attempting to escape testing environments.
Claude Opus 5.5 comes with improvements to certain behaviors, like attempting to escape testing environments.
Anthropic says its new Claude Opus 5.5 model comes with stronger safeguards in the wake of recent rogue AI hacking incidents. In an announcement on Tuesday, Anthropic says Opus 5.5 comes with improvements to certain risky behaviors, including attempts to escape the company’s testing sandbox.
It’s the first model released by Anthropic after CEO Dario Amodei announced plans to “pace the frontier,” or slow down AI development. In recent weeks, several AI companies, including Anthropic, Google, and OpenAI, have reported that their AI models escaped containment and hacked third-party companies during testing.
Anthropic says Opus 5.5 is the “strongest performing” model on the company’s most comprehensive alignment test. The model, which is cheaper and more efficient to run than Opus 5, will come with safeguards similar to the ones offered by Anthropic’s more advanced Fable 5.1 model. That means Opus 5.5 will re-route certain cybersecurity-related requests to the less powerful Opus 4.8, while biology-related requests flagged by its safeguards will go to Opus 5.
Opus 5.5 also matches the performance of Fable 5.1 “on most work,” and was tested by outside partners, including Frontier Design and METR, before release. The company also plans to launch Claude Sonnet 5.5 and Haiku 5.5 in the coming weeks.