Deepseek (shēndù qiúsuô – which literally translated as ‘depth seek search’) is the latest
deep learning Ai (Artificial Intelligence) model coming out of China and taking the the Ai world by storm. Headlines included:
-Insane Performance
-Ai Breakthrough
-Deepseek is Insane
-This is scary
-OpenAI (Chat Gpt) replacement
-Next revolution in coding
-Best LLM (Large Language Model)
-Deepseek shocks Ai industry
Deepseek just revolutionise Ai forever among many other positives that Ai experts and reviewers have attributed to the newly appeared open sourced LLM model by a Chinese Ai company.
The Company itself is fully funded by Chinese hedge fund High-flyer. Both Deepseek and High-flyer are stationed in Hangzhou, Zhejiang. The city known as the ‘House of Silk’ during old times, now houses many of the China and global tech giants. Advanced manufacturing, IT, research, Ai, all happening in this city of 13 million people.
Trump vision of keeping America ahead of China in the Ai world may now be in tatters. Even Eric Schmidt (ex-google CEO) has to revise his statement of last year that China is behind US by at least 2-3 years in the world of Ai. He recently said, after the unveiling of Deepseek, that China did a lot of catchup in the past six months to get to par with the US. Sam Altman, CEO of open Ai, even took a jab at Deepseek, saying, “It is relatively easy to copy something that you know works. It is extremely difficult to do something new, risky, and difficult when you don’t know if it will work”.
The bottom line is that Deepseek is performing at par or exceeding the OpenAI at benchmark tests and that’s when the pricing comes into play. Deepseek model delivers almost the same performance at around 96% discount. ($100 vs $4).
What is Deepseek R1?
Deepseek R1 model is gunning straight for OpenAI’s throne. Imagine getting performance similar to OpenAI’s o1 or even the o1 Pro, but for free (or 4%% the cost for the API). Deepseek lets you run R1 on your own machine or the cloud if you prefer. And yes, Deepseek also have an app version.
Unlike o1, which locks researchers out with a black-box approach, R1 offers transparency. It’s open-source, meaning developers can study, modify, and even fine-tune it freely under an MIT license. This openness has drawn praise from AI experts, many of whom describe it as the next big thing in AI.
While o1 Pro costs a steep $200/month, DeepSeek-R1 provides nearly the same functionality for free—or at an unbelievably low price. This has got a lot of users thinking on whether they should continue or cancel their $200 OpenAI subscriptions.
How deep is your learning?
DeepSeek-R1 excels in multiple domains:
Reasoning and Analytics: Matches OpenAI o1 in chemistry, mathematics, and even coding challenges.Efficient and Runs Locally: Deepseek can run smoothly on devices with limited computational power, such as smartphones or personal laptops, thanks to “distilled” versions of the model. Versatility: Runs both locally and on servers, giving users control over their data and processes. This flexibility extends to features like accessing the web, handling complex tasks, and processing uploaded PDFs, which it performs seamlessly.
However, as with most models, it occasionally hallucinates answers to abstract questions. But its ability to execute practical tasks outweighs these quirks. A couple of reviewers did complain about bias in answers generated out of Deepseek too.Value proposition
One of Deepseek R1’s most game-changing aspects is its cost-effectiveness. For example, tasks that cost $300 on OpenAI o1 can be executed for under $10 with R1. This affordability has made high-performance AI attainable for smaller businesses, researchers, and developers without significant budgets.
DeepSeek’s success is not just about technical breakthroughs but also about overcoming geopolitical challenges. Despite U.S. export restrictions on latest Nvidia advanced AI chips, DeepSeek built R1 efficiently and affordably using older chips. This demonstrates that innovation is not solely dependent on massive budgets or cutting-edge hardware.
The AI community has taken notice, with users worldwide testing and integrating R1. Developers are using it to create tools like Perplexity AI clones, chat apps for PDF analysis, and browser-based AI solutions—all at remarkably lower costs and faster speeds.
Open AI has introduced o3 mini, a new reasoning model, now available for free. Plus-tier subscribers also benefit from extensive o3 mini usage. However, while Deepseek AI offers unlimited access and o3 mini comes with significant limitations. This still makes OpenAI much more expensive than DeepSeek, which remains essentially free.
The potential applications for DeepSeek R1 are limitless. DeepSeek is transforming AI from an elite tool into a universal resource—accessible, affordable, and open for everyone. The question isn’t whether it will change the game. It already has. In a world where AI seems to be getting more expensive and exclusive, DeepSeek feels like a breath of fresh air. Sure, it’s not perfect, but it’s close enough to make the big players nervous. And that’s saying something. From political perspective, it could be the evidence of China rolling out the concept of ‘shared prosperity’ globally.
Let’s use it now!
If you haven’t checked it out yet, go to chat.deepseek.com to try out the depth of its brain. You can also install the Deepseek app available on the App Store and the Google Play Store. It has the logo of a light blue whale.
