360 Security LLM 360安全大模型
Enterprise threat-detection model behind 360's ~1,000 bug finds
Overview
360 Security LLM is Qihoo 360 Digital Security's in-house cybersecurity model family, spanning attack detection, SOC operations, and code auditing. Its vulnerability-discovery agent has found close to 1,000 unknown flaws, including in Microsoft Office and OpenClaw.
What you can produce with 360 Security LLM 360安全大模型
- Six-model security matrix - attack detection, SOC ops, code auditing, threat tracing, knowledge management, data protection
- AI vulnerability-discovery agent - ~1,000 unknown flaws found, incl. Microsoft Office and OpenClaw
- AISF framework spanning model, infrastructure, agent, and application layers
- Security agents certified by China Software Testing Center (Mar 2026)
- Deep integration across 360's endpoint, traffic-detection, and SOC products
ASEAN Perspective
360 Security LLM 360安全大模型 in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
360's security model line is one of the more credible AI-finds-real-bugs stories out of China this year. Its vulnerability-discovery agent, part of a six-model matrix covering attack detection, security operations, code auditing, data protection, and threat tracing, is credited with surfacing close to 1,000 previously unknown vulnerabilities, including issues in Microsoft Office and the open-source OpenClaw agent framework, per Bloomberg's April 2026 reporting. Founder Zhou Hongyi has claimed it lifts a single analyst's output roughly eightfold.
This isn't a product you sign up for on a pricing page — it's baked into 360's own endpoint, traffic-detection, and SOC tooling, sold through an enterprise channel with no public pricing or API. For ASEAN buyers, that means engaging 360 directly and weighing data-sovereignty questions, since the model is trained and operated inside China's security ecosystem. Judge it on capability, not accessibility.
What people say
There's no G2 or Capterra page for 360 Security LLM — it isn't sold as self-serve SaaS, so the record here comes from trade press and 360's own technical disclosures rather than user reviews. That's worth noting upfront: this write-up leans on what journalists and security researchers verified, not on independent customer testimonials.
The strongest independent validation is Bloomberg's April 2026 report that 360 Digital Security's AI vulnerability-discovery agent has found nearly 1,000 previously unknown software flaws, including bugs in Microsoft Office and in OpenClaw, an open-source AI-agent framework. Bloomberg framed this in the same category as Western AI bug-hunting projects, a comparison that signals 360 is being taken seriously by outside observers rather than just touted internally.
360 founder Zhou Hongyi has separately claimed, in an interview with MIT Technology Review China, that the security large model lifts a single security analyst's effective output by roughly eight times, and that existing 360 customers get free upgrades to the new model. Chinese trade coverage (CSDN, Anquanke) describes the underlying architecture as a matrix of six specialized models, covering attack detection, security operations, knowledge management, threat tracing, code security, and data protection, unified under what 360 calls its AI Security Framework (AISF), spanning model, infrastructure, agent, and application layers. In March 2026, 360's security agents reportedly became among the first in China to pass certification from the China Software Testing Center, a formal government-adjacent quality signal rather than a marketing claim.
What's missing from the public record is any account from a named enterprise customer, any comparative benchmark against Western vulnerability-research tools, or any pricing detail — everything is filtered through 360's own announcements or reporters summarizing them. For a domestic Chinese buyer already inside the 360 ecosystem, the vulnerability-discovery numbers are a genuinely strong signal. For anyone else, there's currently no way to independently verify claims beyond what 360 and the outlets covering its announcements have published.
Summary of public user & expert reviews, compiled by RECATOOLS.
About this listing
This entry was compiled from publicly available data including 360 Security LLM 360安全大模型's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with 360 Security LLM 360安全大模型 unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to 360 Security LLM 360安全大模型 directly →
Spotted something out of date? Suggest an update →
360 Security LLM 360安全大模型 in the news
Cybersecurity
Langflow Becomes the First AI-Agent Platform on CISA's Exploited-Vulnerability List
Cybersecurity
CVE-2026-8451: A CitrixBleed-Class NetScaler Flaw Was Exploited Within a Day of Disclosure
Cybersecurity
Google and the FBI Disrupt NetNut, a 2-Million-Device Residential Proxy Network Fuelled by...
More in Security & Safety
