汉字News

← 汉字NewsChinese News for Learners
technology
kē jì
科技
·HSK 6

Anthropic (AI company)
Anthropic
Anthropic
under / affiliated with
qíxià
旗下
new AI model
xīn móxíng
新模型
Mythos (model code name)
Mythos
Mythos
accidentally
yìwài
意外
leaked / disclosed
xièlù
泄露
，
cybersecurity
wǎngluò ānquán
网络安全
experts
zhuānjiā
专家
to issue / to raise
fāchū
发出
unprecedented
qián suǒ wèi yǒu
前所未有
risk
fēngxiǎn
风险
warning
jǐnggào
警告
。

Anthropic (AI company)
Anthropic
Anthropic
due to
yīn
因
data storage
shùjù cúnchǔ
数据存储
oversight / security gap
shūlòu
疏漏
，
accidentally
yìwài
意外
particle introducing the object
jiāng
将
a new
yī kuǎn
一款
performance
xìngnéng
性能
far surpassing
yuǎn chāo
远超
existing
xiànyǒu
现有
models
móxíng
模型
possessive particle
de
的
new type of
xīnxíng
新型
artificial intelligence
réngōng zhìnéng
人工智能
Mythos (AI model)
Mythos
Mythos
possessive particle
de
的
internal documents
nèibù wénjiàn
内部文件
to expose
bàolù
暴露
on / to
yú
于
public
gōngkāi
公开
internet / network
wǎngluò
网络
。
the model
gāi móxíng
该模型
is considered to be
bèi rènwéi
被认为
to possess
jùbèi
具备
unprecedented
qián suǒ wèi yǒu
前所未有
possessive particle
de
的
cyberattack
wǎngluò gōngjī
网络攻击
capability
nénglì
能力
，
the news
xiāoxi
消息
coming to light
bàoguāng
曝光
after
hòu
后
to trigger
yǐnfā
引发
market
shìchǎng
市场
panic
kǒnghuāng
恐慌
，
multiple
duō jiā
多家
cybersecurity
wǎngluò ānquán
网络安全
companies
gōngsī
公司
stock prices
gǔjià
股价
in immediate response
yīng shēng
应声
to fall / to drop
xiàdiē
下跌
。

March 27, 2026 at 8:00 PM

the United States
Měiguó
美国
artificial intelligence
réngōng zhìnéng
人工智能
company
gōngsī
公司
Anthropic (AI company)
Anthropic
Anthropic
due to
yīn
因
operational error / handling mistake
cāozuò shīwù
操作失误
，
particle introducing the object of an action
jiāng
将
a copy of
yī fèn
一份
internal
nèibù
内部
draft document
cǎogǎo
草稿
blog post
bókè wénzhāng
博客文章
accidentally
yìwài
意外
to expose / to make public
bàolù
暴露
in / on
zài
在
publicly accessible
gōngkāi
公开
searchable
kě sōusuǒ
可搜索
modifying particle
de
的
database / data store
shùjùkù
数据库
within
zhōng
中
。
this document
zhè fèn
这份
document / file
wénjiàn
文件
to reveal / to disclose
jiēshì
揭示
completed action particle
le
了
a new
yī kuǎn
一款
code name
dàihào
代号
named / designated as
wéi
为
Mythos (AI model code name)
Mythos
Mythos
modifying particle
de
的
brand new
quánxīn
全新
artificial intelligence
réngōng zhìnéng
人工智能
model
móxíng
模型
，
Anthropic (AI company)
Anthropic
Anthropic
describes it
chēng
称
it / its
qí
其
as
wéi
为
the company
gāi gōngsī
该公司
up to now / to date
qìjīn
迄今
to build / to develop
gòujiàn
构建
modifying particle
de
的
the most powerful
zuì qiángdà
最强大
model
móxíng
模型
。
according to
jù
据
leaked
xièlù
泄露
content
nèiróng
内容
to show / to indicate
xiǎnshì
显示
，
Mythos (AI model)
Mythos
Mythos
in / on
zài
在
software programming
ruǎnjiàn biānchéng
软件编程
、
academic reasoning
xuéshù tuīlǐ
学术推理
and / as well as
jí
及
cybersecurity
wǎngluò ānquán
网络安全
and so on
děng
等
multiple
duō xiàng
多项
benchmarks / tests
cèshì
测试
in / among
zhōng
中
all / uniformly
jūn
均
by a large margin
dàfú
大幅
to surpass
chāoyuè
超越
Claude Opus 4.6 (previous top AI model)
Claude Opus 4.6
Claude Opus 4.6
。
the document
wénjiàn
文件
also / additionally
hái
还
to warn
jǐnggào
警告
that / stating
chēng
称
，
Mythos (AI model)
Mythos
Mythos
in / regarding
zài
在
cyber capabilities
wǎngluò nénglì
网络能力
in terms of
fāngmiàn
方面
currently
mùqián
目前
to be ahead of
lǐngxiān yú
领先于
any
rènhé
任何
other
qítā
其他
artificial intelligence
réngōng zhìnéng
人工智能
model
móxíng
模型
，
potentially
kěnéng
可能
to facilitate / to encourage
zhùzhǎng
助长
large-scale
dàguīmó
大规模
cyberattacks
wǎngluò gōngjī
网络攻击
。
Anthropic (AI company)
Anthropic
Anthropic
spokesperson
fāyán rén
发言人
to confirm
zhèngshí
证实
the model
gāi móxíng
该模型
to represent / to signify
dàibiǎo
代表
the company
gōngsī
公司
artificial intelligence
réngōng zhìnéng
人工智能
performance
xìngnéng
性能
possessive particle
de
的
step-change in capability
jiēyuè shì
阶跃式
breakthrough
tūpò
突破
，
but / however
dàn
但
to emphasize
qiángdiào
强调
currently
mùqián
目前
still
réng
仍
to be in a state of
chǔyú
处于
early-stage
zǎoqī
早期
testing
cèshì
测试
phase / stage
jiēduàn
阶段
，
not yet
shàngwèi
尚未
to / toward
xiàng
向
the public
gōngzhòng
公众
to release
fābù
发布
。
approximately
yuē
约
around 3,000 files
sānqiān fèn
3000份
associated with
yǔ
与
Anthropic (AI company)
Anthropic
Anthropic
blog
bókè
博客
related
xiāngguān
相关
modifying particle
de
的
digital assets / files
shùzì zīchǎn
数字资产
had previously been
céng
曾
stored in
cún yú
存于
unencrypted
wèi jiāmì
未加密
modifying particle
de
的
publicly accessible
gōngkāi
公开
database
shùjùkù
数据库
，
this incident
cǐ cì
此次
information leak
xìnxī xièlù
信息泄露
precisely
zhèng
正
originated from
yuán yú
源于
this
zhè yī
这一
storage
cúnchǔ
存储
oversight / gap
shūlòu
疏漏
。
the company
gōngsī
公司
already / has
yǐ
已
privately
sīxià
私下
to / toward
xiàng
向
government
zhèngfǔ
政府
senior-level / high-ranking
gāocéng
高层
officials
guānyuán
官员
to issue / to send
fāchū
发出
warning
jǐnggào
警告
，
stating that
chēng
称
Mythos (AI model)
Mythos
Mythos
possessive particle
de
的
to be released / to arrive
wènshì
问世
will
jiāng
将
to cause / to make
shǐ
使
the year 2026
èr líng èr liù nián
2026年
to break out
bàofā
爆发
large-scale
dàguīmó
大规模
cyberattacks
wǎngluò gōngjī
网络攻击
possessive particle
de
的
possibility / likelihood
kěnéng xìng
可能性
significantly
dàfú
大幅
to increase / to raise
tíshēng
提升
。
the news
xiāoxi
消息
came to light / was exposed
bàoguāng
曝光
after / following
hòu
后
，
multiple
duō jiā
多家
cybersecurity
wǎngluò ānquán
网络安全
companies
gōngsī
公司
possessive particle
de
的
stock price
gǔjià
股价
immediately
suíjí
随即
to fall / to drop
xiàdiē
下跌
，
market
shìchǎng
市场
regarding
duì
对
artificial intelligence
réngōng zhìnéng
人工智能
models
móxíng
模型
potentially
kěnéng
可能
being used for
bèi yòng yú
被用于
hacker attacks / cyberattacks
hēikè gōngjī
黑客攻击
possessive particle
de
的
concern / worry
dānyōu
担忧
sharply
jíjù
急剧
to heat up / to intensify
shēngwēn
升温
。

News in English

Science & Technology

Anthropic Accidentally Leaks Powerful New AI Model Mythos, Experts Warn of Unprecedented Cybersecurity Risk

Anthropic accidentally exposed internal documents about its new AI model Mythos through a data storage oversight, revealing unprecedented cybersecurity risks; the news rattled markets and sent cybersecurity stocks lower.

More in Science & Technology

Microsoft
Wēiruǎn
微软
and / with
yǔ
与
OpenAI (AI company)
OpenAI
OpenAI
to revise / to amend
xiūɡǎi
修改
cooperation / partnership
hézuò
合作
agreement / terms
xiéyì
协议
，
to give up / to relinquish
fànɡqì
放弃
exclusive
dújiā
独家
license / authorization
shòuquán
授权
。

Microsoft and OpenAI Revise Partnership to End Exclusive Licensing Arrangement

Dell (company)
Dài ěr
戴尔
first quarter
dì yī jìdù
第一季度
AI (artificial intelligence)
AI
AI
server
fúwùqì
服务器
revenue
yínɡshōu
营收
year over year
tónɡbǐ
同比
to surge dramatically
bàozēnɡ
暴增
757 percent
757%
757%
。

Dell AI Server Revenue Surges 757 Percent Year Over Year in Record Quarter

United Kingdom
Yīngguó
英国
researchers
yánjiū rényuán
研究人员
to obtain / to gain access to
huòdé
获得
Google
Gǔgē
谷歌
Willow (Google quantum chip)
Willow
Willow
quantum
liàngzǐ
量子
chip
xīnpiàn
芯片
access rights / right to use
shǐyòngquán
使用权
。

UK Researchers at Kings College London Gain Access to Google Willow Quantum Chip

United States
Měiguó
美国
Illinois (US state)
Yīlìnuòyī Zhōu
伊利诺伊州
to pass (legislation)
tōngɡuò
通过
across the US / nationwide
quánměi
全美
most stringent / strictest
zuì yángé
最严格
AI (artificial intelligence)
AI
AI
safety / security
ānquán
安全
bill / legislation
fǎàn
法案
。

Illinois Passes Nation Most Stringent AI Safety Bill Requiring Third-Party Audits