汉字News

← 汉字NewsChinese News for Learners
technology
kē jì
科技
·HSK 3

OpenAI (AI company)
OpenAI
OpenAI
to release / to launch
tuīchū
推出
GPT-5.5 Instant (AI model)
GPT-5.5 Instant
GPT-5.5 Instant
，
to become
chéngwéi
成为
ChatGPT (AI chatbot)
ChatGPT
ChatGPT
new
xīn
新
default
mòrèn
默认
model
móxíng
模型
。

OpenAI (AI company)
OpenAI
OpenAI
on (a date) / at
yú
于
May 5th
wǔ yuè wǔ rì
五月五日
to release / to launch
tuīchū
推出
GPT-5.5 Instant (AI model)
GPT-5.5 Instant
GPT-5.5 Instant
，
to replace / to supersede
qǔdài
取代
old / previous
jiù
旧
version
bǎnběn
版本
to become
chéngwéi
成为
ChatGPT (AI chatbot)
ChatGPT
ChatGPT
structural particle (possessive)
de
的
new
xīn
新
default
mòrèn
默认
model
móxíng
模型
。
the said / this
gāi
该
model
móxíng
模型
significantly / by a large margin
dàfú
大幅
to reduce / to decrease
jiǎnshǎo
减少
errors / mistakes
cuòwù
错误
，
and / also
bìng
并
to enhance / to strengthen
zēngqiáng
增强
personalization
gèxìnghuà
个性化
features / functions
gōngnéng
功能
。

May 7, 2026 at 7:00 AM

OpenAI (AI company)
OpenAI
OpenAI
at / on
zài
在
May 5th
wǔ yuè wǔ rì
五月五日
to begin / to start
kāishǐ
开始
toward / to
xiàng
向
all / every
suǒyǒu
所有
users
yònghù
用户
to push out / to roll out
tuīsòng
推送
GPT-5.5 Instant (AI model)
GPT-5.5 Instant
GPT-5.5 Instant
。
this / this one
zhège
这个
new
xīn
新
model
móxíng
模型
to replace
qǔdài
取代
GPT-5.3 Instant (previous model)
GPT-5.3 Instant
GPT-5.3 Instant
，
to become
chéngwéi
成为
ChatGPT (AI chatbot)
ChatGPT
ChatGPT
structural particle (possessive)
de
的
default
mòrèn
默认
model
móxíng
模型
。
according to / based on
gēnjù
根据
OpenAI (AI company)
OpenAI
OpenAI
structural particle (possessive)
de
的
internal / in-house
nèibù
内部
evaluation / assessment
pínɡcè
评测
，
GPT-5.5 Instant (AI model)
GPT-5.5 Instant
GPT-5.5 Instant
produced / generated
chǎnshēng de
产生的
hallucinations (AI errors)
huànjué
幻觉
compared to
bǐ
比
the previous generation
shàng yīdài
上一代
fewer / less
shǎo
少
52.5 percent
52.5%
52.5%
。
answers / responses
huídá
回答
structural particle
de
的
word count / length
zìshù
字数
also
yě
也
to decrease / to reduce
jiǎnshǎo
减少
completed action particle
le
了
approximately
yuē
约
30 percent
30%
30%
，
tone / style
yǔqì
语气
more / even more
gèng
更
natural / casual
zìrán
自然
。
new
xīn
新
model
móxíng
模型
also / in addition
hái
还
to support
zhīchí
支持
personalization
gèxìnghuà
个性化
features
gōngnéng
功能
，
can / is able to
kěyǐ
可以
to refer to / to consult
cānkǎo
参考
user
yònghù
用户
structural particle
de
的
history / past
lìshǐ
历史
conversations
duìhuà
对话
and
hé
和
files / documents
wénjiàn
文件
。
this
zhè yī
这一
feature
gōngnéng
功能
currently / at present
mùqián
目前
to / toward
xiàng
向
Plus (subscription tier)
Plus
Plus
and
hé
和
Pro (subscription tier)
Pro
Pro
users
yònghù
用户
to open / to make available
kāifàng
开放
。
developers
kāifāzhě
开发者
can / are able to
kěyǐ
可以
through / via
tōngguò
通过
API (application programming interface)
API
API
using / with
yǐ
以
chat-latest (API model name)
chat-latest
chat-latest
to call / to access
diàoyòng
调用
the said
gāi
该
model
móxíng
模型
。

New Words

tuīsòng
推送
to roll out / to push
qǔdài
取代
to replace
mòrèn
默认
default
huànjué
幻觉
hallucinations
jiǎnshǎo
减少
to reduce / to decrease
zìrán
自然
natural
gèxìnghuà
个性化
personalization
cānkǎo
参考
to refer to / to consult
duìhuà
对话
conversation
diàoyòng
调用
to call / to access

News in English

Science & Technology

OpenAI Releases GPT-5.5 Instant as New Default ChatGPT Model

OpenAI launched GPT-5.5 Instant on May 5 as the new default model in ChatGPT, replacing GPT-5.3 Instant with significantly fewer hallucinations and improved personalization.

OpenAI began rolling out GPT-5.5 Instant to all users starting May 5th. This new model replaces GPT-5.3 Instant and becomes the default model for ChatGPT.

According to OpenAI's internal evaluations, GPT-5.5 Instant produces 52.5% fewer hallucinations compared to the previous generation. The word count of responses has also decreased by approximately 30%, with a more natural and casual tone.

The new model also supports personalization features that can reference user conversation history and documents. This feature is currently available to Plus and Pro users. Developers can call this model through the API using the chat-latest endpoint.

More in Science & Technology

Microsoft
Wēiruǎn
微软
to launch / to release
tuīchū
推出
Work IQ (Microsoft enterprise AI tool)
Work IQ
Work IQ
，
to let / to allow
ràng
让
Copilot (Microsoft AI assistant)
Copilot
Copilot
to connect to
liánjiē
连接
enterprise data
qǐyè shùjù
企业数据
。

Microsoft Launches Work IQ, Letting Copilot Connect to Enterprise Data

Nous Research (AI research company)
Nous Research
Nous Research
to complete
wánchéng
完成
$90 million
jiǔ qiān wàn měiyuán
九千万美元
to raise funding
róngzī
融资
，
Hermes (open-source AI model)
Hermes
Hermes
to advance into / to enter
jìnjūn
进军
enterprise market
qǐyè shìchǎng
企业市场
。

Nous Research Raises $90 Million, Hermes Enters Enterprise Market

former / previous
qián
前
Groq (AI chip startup)
Groq
Groq
engineers
gōngchéngshī
工程师
regarding / over (a matter)
jiù
就
$20 billion USD
200 yì měiyuán
200亿美元
Nvidia (US semiconductor company)
Yīngwěidá
英伟达
licensing / authorization
shòuquán
授权
deal / transaction
jiāoyì
交易
to file / to initiate
tíqǐ
提起
a lawsuit / legal action
sùsòng
诉讼
。

Former Groq Engineers Sue Board Over $20 Billion Nvidia Licensing Deal

Qualcomm (US semiconductor company)
Gāotōng
高通
for the first time
shǒucì
首次
to / toward
xiàng
向
Huawei (Chinese tech company)
Huáwéi
华为
to pay / to make a payment
zhīfù
支付
patent royalties / licensing fee
zhuānlì fèi
专利费
，
both parties
shuāngfāng
双方
to sign / to conclude
qiānshǔ
签署
multi-year period
duō nián qī
多年期
agreement / deal
xiéyì
协议
。

Qualcomm Pays Huawei for First Time in Broad Multi-Year Patent Licensing Agreement