DeepSeek, China's AI model: News & Discussion
- Thread starter Beijingwalker
- Start date
F-22Raptor
Elite Member
The Chinese and their sycophants couldn’t stop yapping about Chinese efficiency a few days ago and how inefficient US AI companies were.
OpenAI crushed that false narrative with o3
UKBengali
Elite Member
The Chinese and their sycophants couldn’t stop yapping about Chinese efficiency a few days ago and how inefficient US AI companies were.
OpenAI crushed that false narrative with o3
O3 mini is barely better than R1.
F-22Raptor
Elite Member
It’s better, faster, and nearly as efficient.
It’s only the mini version which is more focused on specific tasks like coding.
o3 full and pro versions expected in 4-6 weeks.
F-22Raptor
Elite Member
F-22Raptor
Elite Member
F-22Raptor
Elite Member
UKBengali
Elite Member
Still not much better.
Wait and see for full and pro versions of o3.
This was Deepseek’s first go and they are working on more advanced models as we speak.
F-22Raptor
Elite Member
The narrative last week was how China had surpassed the US in AI and how the US could never compete on efficiency.
OpenAI crushed that narrative in the matter of days.
UKBengali
Elite Member
Nope as we have no data for efficiency as in how much it cost exactly to train this new model.
Remember Deepseek started from zero and so of course they have a disadvantage whereas OpenAI were able to build on what they had done before.
Deepseek are now using far more powerful Huawei Ascend 910C and so they will more than likely have a new "surprise" to send Nvidia and the other US bloated tech stocks crashing down again.
OpenAI is now following leader in AI Deepseek, which is a Chinese company of just 200 employees and their AI talent is graduates and those still doing their PhDs.
F-22Raptor
Elite Member
F-22Raptor
Elite Member
Also Sam Altman confirmed GPT-5 will reach Level 4 on the AGI scale
UKBengali
Elite Member
Total cope, take your L
Ok, managed to get some actual data at last. Still twice as expensive as Deepseek R1 for comparable performance.
Massive improvement but still nowhere near the level of Deepseek's model.
OpenAI releases its new o3-mini reasoning model for free
OpenAI just released o3-mini, a reasoning model that’s faster, cheaper, and more accurate than its predecessor.
www.technologyreview.com
The company says its new model, o3-mini, costs 63% less than o1-mini per input token However, at $1.10 per million input tokens, it is still about seven times more expensive to run than GPT-4o mini.
Additionally, DeepSeek’s reasoning model costs $0.55 per million input tokens, half the price of o3-mini, so OpenAI still has a way to go to bring down its costs. It’s estimated that reasoning models also have much higher energy costs than other types, given the larger number of computations they require to produce an answer.
Pakistan Space Agency
Elite Member
Britain is Great Britain. The Conqueror of Empires from East to West.
Pakistan Space Agency
Elite Member
Amazing. I thought it was CGI at first.
DJI Sky City, the global headquarters of the world's leading consumer drone marker DJI, in Shenzhen.
Users who are viewing this thread
Pakistan Defence Latest
Country Watch Latest
-
Iran - Israel/US War: Israel-US declare war on Iran, Iran responds (97 Viewers)- Latest: Persian Gulf
-
-
-
-
Latest Posts
-
Iran - Israel/US War: Israel-US declare war on Iran, Iran responds (97 Viewers)
- Latest: Persian Gulf
-
-
-
