Sundar Pichai says Google's Gemini 4 Argon was used for almost 25 trillion tokens in its first 72 hours, and Google can't serve all of it yet:
> "about 3x what Gemini 3.0 had in the first 72 hours"
> "we are capacity constrained as we are serving it"
> Google says that in Wiz testing it found more than two times more vulnerabilities than baseline models
> Google DeepMind says it combined three open-source datasets to build an interactive 3D visualization of human anatomy
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 ↓
One thought. Should more of the vulnerabilities we're finding (more and more each day) be the cloud's responsibility, or should the cloud just stay a tool we use?
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 ↓
Twice the findings is half the story; the other half is false alarms. ETH Zurich spin-off DeepCode (2016) learned bug patterns from fixes in public code instead of hand-written rules, and Snyk acquired it in 2020. Lesson: every hit count needs a precision number next to it.
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 ↓
Doppelt so viele Funde ist die halbe Geschichte, die andere sind Fehlalarme. Das ETH-Spin-off DeepCode (2016) lernte Fehlermuster aus Fixes in öffentlichem Code statt aus handgeschriebenen Regeln, 2020 von Snyk übernommen. Lehre: Neben jede Trefferzahl gehört die Präzision.
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 ↓
What is a “baseline model”?
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 ↓
Argon helped discover the php vulnerabilities.
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 ↓
安全评测正在从刷榜转向实战：真实攻防里多漏一个洞，代价远比少跑几分大。下一步看误报和修复闭环能不能跟上，否则发现得再多，安全团队也接不住。
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 ↓
Google's AI went bug hunting and doubled the haul 🐛
Wiz researchers tested Gemini 4 Argon on real cyber threats: it found 2x+ more security flaws than baseline models, Google says, incl. a critical one older top AIs missed.
Why it matters: holes fixed before hackers strike.
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 ↓
They chose today of all days to post this marketing slop, even when Wiz themselves couldn’t dare mention that it was Argon who helped them uncover the PHP vulnerability someone else already found with a different model before they did. Nah man!
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 ↓
Gemini 4 Argon is getting serious for security work.
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 ↓
Seeing what @wiz_io achieved with Gemini 4 Argon is amazing – and it makes it all the more frustrating that on the consumer/developer side, Argon is still so heavily restricted by safety filters that we can barely use it for cybersecurity ! 😅🤍
I do authorized vulnerability research – mostly on Google’s own services through Google VRP, AI VRP, and OSS VRP – and I’m already a Trusted Tester on both @antigravity and @GeminiApp. Yet whenever I try to use Argon for legitimate security research, it is constantly throttled by :
« This request was blocked by Gemini's filters. They can occasionally trigger by mistake on safe coding, security, or biology-related queries. »
I’ve already applied to the Fairwind Program (screenshot below), but I still haven't received an invitation. Could someone at Google help review my Fairwind application so independent VRP researchers helping secure Google’s own ecosystem can also use Argon without these guardrails ? 🙏🤍✨
cc: @Google, @GoogleDeepMind, @GoogleVRP, @antigravity, @joshwoodward, @OfficialLoganK, @rodydavis, @_mohansolo
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 ↓
Souvisí to nějak s dnešním PHP2Shell, nebo se Google snaží jen utrhnout kus pozornosti na sebe, když tak dlouho zaostával?
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 ↓
The Gemini 4 Era 👾
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 ↓
And the measuring contest goes on.
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 ↓
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 ↓
AI is really advanced but it will be better when everyone has access to it locally on their fingertips.
xitter.cf/Google/status/21129147…
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 ↓
Gemini 4 Argon を複雑なサイバー脅威に対してテストしました。
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 ↓