Models

OpenAI 发布 GPT-6.2 Astra:网络攻防能力提升 22 分,同步收紧安全护栏

Why we picked it

OpenAI 把网络攻防能力当卖点来发版,ExploitBench 涨 22 分的同时承诺收紧护栏。如果你做安全研究或红队工作,能不能进 Daybreak 拿到低拒绝率访问才是关键信息,值得看看审核机制怎么设计。

OpenAI 发布 GPT-6.2 Astra,称其为迄今网络攻防能力最强的模型。在内部基准 ExploitBench Internal Port 上,GPT-6.2 Astra 较上一代 GPT-6 Astra 提升 22 个百分点。

Body

由于能力增强带来的滥用风险,OpenAI 同步为 GPT-6.2 Astra 和 GPT-6.2 Sol 配套了更严格的安全护栏,对高风险网络安全协助请求加以限制。同时,OpenAI 正在推进将这两个模型接入 Daybreak 计划,让通过审核的防御方研究人员以更少的拒绝率访问模型能力。

这一「能力提升与护栏收紧」同步发布的做法,延续了前沿实验室在网络能力上攻防两线并行的策略。对安全厂商和红队从业者而言,能否进入 Daybreak 审核名单将直接影响其对模型利用能力的实际获取程度。

Original · OpenAI

GPT-6.2 Astra is our most capable model ever for cyber. It jumps 22 points over GPT-6 Astra on ExploitBench Internal Port. Due to increased capabilities, we’re launching GPT‑6.2 Astra and Sol with tighter safeguards for higher-risk cybersecurity assistance. We are working to bring these models to Daybreak so approved defenders can access them with reduced refusals.📊 9 ⚡

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include GPT-6.2 Astra, OpenAI, cybersecurity, reasoning models.

Source attribution: curated from OpenAI (@OpenAI).