- DeepSeek AI is an AI-powered chatbot similar to ChatGPT, and it has been developed by a Chinese company.
- In early 2025, DeepSeek released its groundbreaking R1 reasoning model, nearly matching OpenAI's o1 reasoning model.
- DeepSeek is currently training its next R2 reasoning model which may arrive in the next few months.
After the ChatGPT moment of 2022, the next major high came in early 2025 with DeepSeek. This relatively unknown company from China entered the AI race and challenged many established players. If you are still wondering what is DeepSeek AI and what it does, we have compiled all the information in one place. On that note, let’s go ahead and learn about DeepSeek AI in detail.
What is DeepSeek AI?
DeepSeek AI is an AI chatbot similar to ChatGPT, and it has been developed by a Chinese company headquartered in Hangzhou. It’s formally registered as Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd. and it’s owned by a Chinese hedge fund High-Flyer. DeepSeek was founded in July 2023 by Liang Wenfeng who is the CEO and co-founder of both High-Flyer and DeepSeek.
Think of DeepSeek AI as a rising rival to Western AI labs like OpenAI and Anthropic. It’s a young startup, but in just two years, it has surprised the world with powerful AI models. DeepSeek AI rose to fame when it launched the R1 reasoning model in January 2025, nearly matching OpenAI’s o1 performance, and surpassed ChatGPT to take the top spot on the US App Store.
What made DeepSeek even more compelling was that it reportedly trained the R1 model under just $6 million, significantly lower than the estimated $100 million required to train Western AI models of similar size. This led to an 18% drop in Nvidia’s share price. Not just that, DeepSeek also provided the model weights publicly, in a push toward open-source AI development.
In addition, DeepSeek released a detailed technical paper and openly shared its RL-based post-training method to train reasoning models. While OpenAI and other Western AI labs closely guarded their RL-based methods to train reasoning models, DeepSeek, a young AI startup from China entered the AI race pretty late and proved the world that powerful reasoning models could be trained at much lower costs.
At the time of the release, OpenAI was asking users to pay $20 to access its o1 reasoning model. However, DeepSeek R1 was released for free, and that pressured AI players including OpenAI to make their advanced AI models available to users at no cost. In a way, China’s DeepSeek broke the Western monopoly on AI and made it more accessible to users worldwide.
AI Models Released by DeepSeek
In 2023, DeepSeek released its first AI model called DeepSeek Coder for coding tasks. In addition, it also released the DeepSeek LLM and DeepSeek-Math models. After just six months, the company announced a much larger DeepSeek-V2 and DeepSeek-Coder V2 AI models.
Later, DeepSeek-V2.5 was released in September. The first reasoning AI model, DeepSeek-R1-Lite was released in November, and in December, we got the DeepSeek-V3 base model. Now, in January 2025, the capable DeepSeek-R1 reasoning model was released with DeepSeek app for both Android and iOS.
In 2025, the company keep updating its DeepSeek-V3 and DeepSeek-R1 models. Now, the DeepSeek-R1-0528 and DeepSeek-V3.2 are the latest versions which were released in September. It’s expected that DeepSeek may release a big AI model in the upcoming months.
How to Use DeepSeek AI?
You can use DeepSeek AI for free via chat.deepseek.com or you can install the DeepSeek mobile app (Android / iOS). Apart from that, you can run DeepSeek R1 locally on your PC or mobile device. That said, you can only download and run distilled versions of the larger model on your local device.
What is Happening with DeepSeek Now?
While DeepSeek was rapidly launching competitive AI models, in recent months, the company has not done any major launches. The reason is that the Chinese government urged DeepSeek to use domestic AI chips like the Huawei Ascend for its follow-up AI model.
The Financial Times reported that Chinese authorities asked DeepSeek to adopt Huawei’s Ascend chips to train the R2 model. As a result, the new model has been delayed. New reports suggest that the DeepSeek team tried to use Ascend chips but switched back to Nvidia GPUs for training due to technical limitations.
Reportedly, DeepSeek is using Ascend chips for inference only. All that said, DeepSeek is a major AI company to come out of China, along with Alibaba, and we should monitor its progress.
