汉字News

← 汉字NewsChinese News for Learners
technology
kē jì
科技
·HSK 5

Tencent (Chinese tech company)
téngxùn
腾讯
to release / to announce
fābù
发布
open-source
kāiyuán
开源
large language model / LLM
dà móxíng
大模型
Hunyuan 3 (Tencent AI model)
hùnyuán sān
混元3
preview version / beta release
yùlǎn bǎn
预览版
。

Tencent Holdings (parent company)
téngxùn kònggǔ
腾讯控股
at / on (formal preposition)
yú
于
April 24th
sì yuè èrshísì rì
4月24日
to release / to announce
fābù
发布
Hunyuan 3 (Tencent AI model)
hùnyuán sān
混元3
preview version / beta release
yùlǎn bǎn
预览版
，
this is
zhè shì
这是
a / one (product)
yī kuǎn
一款
to adopt / to use
cǎiyòng
采用
mixture-of-experts
hùnhé zhuānjiā
混合专家
architecture / framework
jiàgòu
架构
(structural particle)
de
的
large language model
dàxíng yǔyán móxíng
大型语言模型
，
total number of parameters
zǒng cānshù liàng
总参数量
reaching / amounting to
dá
达
295 billion
2950 yì
2950亿
，
to support
zhīchí
支持
256,000 (tokens)
25.6 wàn gè
25.6万个
token (AI term)
lìng pái
令牌
(structural particle)
de
的
context
shàng xià wén
上下文
window
chuāngkǒu
窗口
，
in / by means of
yǐ
以
open-source
kāiyuán
开源
way / approach
fāngshì
方式
to provide / to offer
tígōng
提供
。

April 25, 2026 at 10:00 AM

Tencent Holdings (parent company)
téngxùn kònggǔ
腾讯控股
at / on (formal preposition)
yú
于
April 24th
sì yuè èrshísì rì
4月24日
officially / formally
zhèngshì
正式
to release / to announce
fābù
发布
Hunyuan 3 (Tencent AI model)
hùnyuán sān
混元3
preview version / beta release
yùlǎn bǎn
预览版
，
this is
zhè shì
这是
company
gōngsī
公司
new management / new leadership
xīn guǎnlǐ céng
新管理层
to take office / to assume the role
shàngrèn
上任
since / ever since
yǐlái
以来
to launch / to introduce
tuīchū
推出
(structural particle)
de
的
first / inaugural (product)
shǒu kuǎn
首款
heavyweight / major
zhòngliàngjí
重量级
AI (artificial intelligence)
AI
AI
model / AI model
móxíng
模型
。
Hunyuan 3 (Tencent AI model)
hùnyuán sān
混元3
to adopt / to use
cǎiyòng
采用
mixture-of-experts
hùnhé zhuānjiā
混合专家
architecture / framework
jiàgòu
架构
，
total number of parameters
zǒng cānshù liàng
总参数量
reaching / amounting to
dá
达
295 billion (parameters)
2950 yì gè
2950亿个
，
among which / of which
qízhōng
其中
each time / per inference
měi cì
每次
inference / reasoning
tuīlǐ
推理
only / merely
jǐn
仅
to activate
jīhuó
激活
approximately / about
yuē
约
21 billion (active parameters)
210 yì gè
210亿个
parameters
cānshù
参数
。
this model / the model
gāi móxíng
该模型
to support
zhīchí
支持
at most / up to
zuì duō
最多
256,000 (tokens)
25.6 wàn gè
25.6万个
token (AI term)
lìng pái
令牌
(structural particle)
de
的
ultra-long / extremely long
chāo cháng
超长
context
shàng xià wén
上下文
window
chuāngkǒu
窗口
，
suitable for / applicable to
shìyòng yú
适用于
to process / to handle
chǔlǐ
处理
complex / complicated
fùzá
复杂
(structural particle)
de
的
long documents
cháng wéndàng
长文档
task / assignment
rènwù
任务
。
Tencent (Chinese tech company)
téngxùn
腾讯
in / by means of
yǐ
以
completely / fully
wánquán
完全
open-source
kāiyuán
开源
(structural particle)
de
的
form / format
xíngshì
形式
to release / to publish
fābù
发布
Hunyuan 3 (Tencent AI model)
hùnyuán sān
混元3
，
developers
kāifā zhě
开发者
can / may
kě
可
on / at
zài
在
Hugging Face (AI model platform)
Hugging Face
Hugging Face
and other / such as
děng
等
platform
píngtái
平台
on / above
shàng
上
free of charge / at no cost
miǎn fèi
免费
to obtain / to access
huòqǔ
获取
model / AI model
móxíng
模型
model weights / parameters
quánzhòng
权重
。
according to / based on
jù
据
Tencent (Chinese tech company)
téngxùn
腾讯
to release / to publish publicly
gōngbù
公布
(structural particle)
de
的
test data / benchmark data
cèshì shùjù
测试数据
，
Hunyuan 3 (Tencent AI model)
hùnyuán sān
混元3
in / at
zài
在
multiple / various
duō xiàng
多项
industry
hángyè
行业
benchmark evaluation / benchmark test
jīzhǔn píngcè
基准评测
in / among
zhōng
中
to perform excellently / to stand out
biǎoxiàn chūsè
表现出色
，
especially / in particular
yóuqí shì
尤其是
in / at
zài
在
code generation
dàimǎ shēngchéng
代码生成
and
hé
和
multi-modal
duō mótài
多模态
understanding / comprehension
lǐjiě
理解
field / domain
lǐngyù
领域
。
this time / on this occasion
cǐ cì
此次
to release / to announce
fābù
发布
marks / signifies
biāozhì zhe
标志着
Tencent (Chinese tech company)
téngxùn
腾讯
in / at
zài
在
open-source
kāiyuán
开源
AI (artificial intelligence)
AI
AI
competition / competitive landscape
jìngzhēng
竞争
field / domain
lǐngyù
领域
to continue / ongoing
chíxù
持续
to increase / to step up
jiādà
加大
investment / commitment
tóurù
投入
，
with / against
yǔ
与
domestic and international / within China and abroad
guónèi wài
国内外
multiple / several
duō jiā
多家
tech giants
kējì jùtóu
科技巨头
to launch / to begin
zhǎnkāi
展开
direct competition / head-on competition
zhèngmiàn jìngzhēng
正面竞争
。

New Words

téngxùn
腾讯
Tencent (Chinese tech company)
hùnyuán
混元
Hunyuan (AI model name)
fābù
发布
to release / to announce
móxíng
模型
AI model
jiàgòu
架构
architecture / framework
cānshù
参数
parameters
tuīlǐ
推理
inference / reasoning
kāiyuán
开源
open-source
jīzhǔn píngcè
基准评测
benchmark evaluation
dàimǎ shēngchéng
代码生成
code generation

News in English

Science & Technology

Tencent Releases Open-Source Large Language Model Hunyuan 3 Preview

Tencent Holdings released Hunyuan 3 Preview on April 24, a large language model using a mixture-of-experts architecture with 295 billion total parameters and a 256,000-token context window, made available as open-source.

On April 24, Tencent Holdings officially released a preview version of Hunyuan 3, marking the first major AI model launched by the company's new management team since they took office.

Hunyuan 3 employs a mixture-of-experts architecture with a total of 295 billion parameters. During each inference operation, only approximately 21 billion parameters are activated, making it efficient for processing tasks.

The model supports ultra-long context windows of up to 256,000 tokens, making it suitable for processing complex long-document tasks.

Tencent released Hunyuan 3 in fully open-source form, allowing developers to freely access the model weights on platforms such as Hugging Face.

According to benchmark data released by Tencent, Hunyuan 3 demonstrates excellent performance across multiple industry benchmarks, particularly excelling in code generation and multi-modal understanding.

This release marks Tencent's sustained commitment to increasing investment in the open-source AI competitive landscape, engaging in direct competition with multiple domestic and international tech giants.

More in Science & Technology

MiniMax
MiniMax
MiniMax
releases
fābu
发布
M3
M3
M3
multimodal
duōmótài
多模态
model
móxíng
模型
。

MiniMax M3 Multimodal Model Delivers 9x Faster Performance

Nvidia
yīngwěidá
英伟达
releases
fābu
发布
Cosmos
Cosmos
Cosmos
3
3
3
open-source
kāiyuán
开源
physical
wùlǐ
物理
AI
réngōng zhìnéng
人工智能
model
móxíng
模型
。

Nvidia Releases Cosmos 3 Open-Source Physical AI Model

Apple (company)
Píngguǒ
苹果
to release / to launch
fābù
发布
brand new / completely new
quánxīn
全新
Siri (Apple voice assistant)
Siri
Siri
，
features / functions
gōngnéng
功能
greatly / significantly
dàfú
大幅
to upgrade / upgraded
shēngjí
升级
。

Apple Unveils Completely Rebuilt Siri at WWDC 2026

Germany
Déguó
德国
robot / robotics
jīqìrén
机器人
company
gōngsī
公司
Neura Robotics (German humanoid robot company)
Neura Robotics
Neura Robotics
to complete / to close
wánchéng
完成
$1.4 billion
14 yì měiyuán
14亿美元
funding round
róngzī
融资
。

German Robotics Firm Neura Robotics Closes $1.4 Billion Funding Round