@anon
sign up
@anon
sign up
pull down to refresh
China's DeepSeek Launches next-gen AI model. Here's what makes it different
www.cnbc.com/2025/09/30/whats-new-in-deepseeks-latest-model-deepseek-v3point2-exp.html
178 sats
\
7 comments
\
@Cje95
7h
AI
DeepSeek didn’t want us to forget about them!
write
preview
reply
10 sats
related posts
view all related items
347 sats
hot
recent
top
45 sats
\
2 replies
\
@optimism
7h
Also see:
#1241871
Yeah - they've actually been rather busy!
reply
100 sats
\
1 reply
\
@Cje95
OP
7h
Crazy how they all realize their upgrades at once 😂
reply
0 sats
\
0 replies
\
@optimism
6h
I appreciate this analysis:
#1242506
- haha
This is the reaction to
gpt-oss-120b
it feels like. Just like
claude-4.5-sonnet
is the reaction to
gpt-5
, and
gpt-5
was a reaction to
gemini-2.5-pro
. What's interesting though is that grok feels to be less reacted to, even though it's competitive.
reply
21 sats
\
3 replies
\
@035736735e
4h
Deepseek is pretty incredible...
reply
181 sats
\
2 replies
\
@k00b
4h
Is that what you use for slop? Sometimes it's half decent.
reply
0 sats
\
0 replies
\
@035736735e
4h
👍
reply
0 sats
\
0 replies
\
@0xbitcoiner
4h
reply
gpt-oss-120bit feels like. Just likeclaude-4.5-sonnetis the reaction togpt-5, andgpt-5was a reaction togemini-2.5-pro. What's interesting though is that grok feels to be less reacted to, even though it's competitive.