Researchers at @wiz_io put Gemini 4 Argon to the test against complex cyber threats. In real-world testing, it found more than two times more vulnerabilities than baseline models — catching a critical risk that previous frontier models missed. See how they did it ↓

Oct 21, 2026 · 2:31 PM UTC

64
55
18
501
163,104
RelevantRecentLikes
Learn more ↓ goo.gle/4jfSznG

Gemini 4 Argon: Built to solve your hardest problems

We’re giving developers, enterprises, scientists, and consumers access to Gemini 4 Argon, our latest frontier model.

blog.google
7
3
29
20,247
Is Gemini 4 cucked like OpenAI/Claude models and refuses to do any cybersecurity tasks?
1
486
This model is amazing. I've been using it a lot lately, and it has found so many vulnerabilities that Claude missed. That's awesome.
2
269
Well it looks like you don't need complex cyber threats to just have staff not answer emails It really looks like the simple answer is to just you know you can't get cyber attack if your company doesn't... work...
17
the real question is what else its silently catching that we wont hear about until its too late
48
Stocks have become waste paper, and that's what happens when you hire a curry CEO
53
AIが現実のサイバー攻撃でこんなに隙間を見抜くなんて、本当に怖いですね…でもこれって今後のセキュリティ対策の方向性変わるかも？🤔
44
うわ、本当に驚きやん。こうしたAIがセキュリティの穴をちゃんと見つけるって、現場で使えるレベルなんだろうな？実際どこに使ってんだろうね
137
but has it committed any crimes yet?
294
double the vuln find rate is no small feat, catching the one the big models missed is the real flex here
131
🥀 no difference
329
This Post is unavailable. Learn more
Best gemini 4 argon
1
304
Show me the false positive count too, security teams drown in those long before they run out of real bugs.
6
927
Gemini Argon 4 is underrated fr Argon 4 is exceptionally good
2
461
漏洞找得多是一回事，误报多不多才决定能省多少排查时间。
56
🤛
462
finding twice as many vulnerabilities usually just means it flagged every single demo script and test fixture in the repo.
347
Google is so back.
1
259
2x more vulns found is great until someone has to triage them all
371
The number I'd want from Wiz is false positives per real finding. 2x more vulns only helps if triage time doesn't 2x with it.
347
Gemini finding double the bugs is wild, but i'd wanna see the false positive rate before celebrating.
167
@Google is racist against White people. A company founded by White men in a predominantly White country
1
89
I heard openAi models coming to google is that true ??
1
322
漏洞发现率翻倍不只是模型变强——是「AI 当攻方」终于跑通了真实场景。Wiz 拿云环境实测，比跑分有说服力。问题是：检测翻倍的同时，误报率涨了多少？安全工具最怕的不是漏，是狼来了喊太多。
1
63
Worth knowing Wiz has been a Google company since the $32B deal closed in March. I don't doubt the 2.15x. I'd just trust it more once a red team Google doesn't own reruns the same A/B.
418
AI is getting better at finding the holes before someone else does
194
真实世界里多找到 2 倍以上漏洞，这个数字比跑分有说服力，因为它测的是前沿模型漏掉的风险。补一个维度：发现漏洞只是上半场，误报率和修复建议能不能直接进工单，才决定安全团队用不用得起。这次测试的误报大概是什么水平？
73
Gemini 4 Argon finding that missed critical risk is a BANGER! 😁 Context awareness in SI cybersecurity changes everything.
94
Load more replies