DeepSeek Has Surpassed ChatGPT – A Detailed Look at the Chinese Chatbot

Artificial intelligence (AI) has already become a part of our daily lives. For a long time, ChatGPT from OpenAI was the leader in this field. But now, the situation has changed. A new player – DeepSeek – has caused a real sensation. It has surpassed ChatGPT in popularity in the U.S., the U.K., and China. Its free app has topped the download charts in the App Store, and Silicon Valley is worried. What kind of model is this? How does it work? Why has it become so successful? Let’s dive in.

What Is DeepSeek?

DeepSeek is a new chatbot developed by a Chinese startup of the same name. Its model, DeepSeek-V3, is open-source, allowing researchers and developers to analyze its functionality.

It Can:

• Write code;
• Solve math problems;
• Analyze texts;
• Answer questions;
• Tell jokes.

The main highlight – it’s free. This has made it incredibly popular.

The new AI chatbot was developed by a team led by Liang Wenfeng. They started working on it before U.S. sanctions were imposed, purchasing powerful Nvidia graphics processors. This allowed them to create a competitive AI model for a relatively low cost – only $6 million.

DeepSeek vs. ChatGPT Comparison

When a new chatbot appears, everyone naturally compares it to ChatGPT. Right away, let’s say that the comparison isn’t always in ChatGPT’s favor. But DeepSeek isn’t perfect either:

• Technical Aspects. DeepSeek has a modern architecture similar to GPT-4 but uses fewer resources. OpenAI spent billions of dollars on development, whereas the Chinese team achieved similar results for much less money.
• Functionality. Both are quite similar in capabilities, but there are differences. They can both code and solve problems. ChatGPT is undoubtedly better at creative tasks like writing stories or poetry. DeepSeek, however, works quickly and is free, giving it an edge.
• Limitations and Censorship. ChatGPT filters queries and may refuse to answer certain topics. DeepSeek also has censorship, but its boundaries are not yet fully understood.

Is DeepSeek truly better than ChatGPT? It’s a tough question – especially considering that its servers have been overloaded for the past 24 hours. Can an average user fairly evaluate the chatbot’s advantages in such conditions? Probably not.

Why Did DeepSeek Become So Popular?

DeepSeek has rapidly gained popularity, surpassing ChatGPT in download rankings. Several factors contributed to this:

• Free Access. Most powerful AI models either require a paid subscription or limit free access. DeepSeek, on the other hand, offers a fairly high level of functionality for free.
• Speed. It responds to queries quickly, especially for complex calculations or coding tasks.
• Open Source. Unlike OpenAI, the Chinese company decided to make its model open-source. This allows researchers and developers to improve it, fostering rapid advancement.
• Popularity in China. Due to restrictions on ChatGPT in China, the new project quickly filled the niche among local users before expanding globally.

OpenAI’s and Competitors’ Reactions

DeepSeek has become a serious competitor to OpenAI, Google, and Anthropic. OpenAI accused the Chinese developers of allegedly using its technology to create their own AI model, according to Bloomberg.

Journalists report that Microsoft and OpenAI have already launched an investigation. Last fall, they noticed that individuals linked to the Chinese project were mass-extracting data from one of OpenAI’s research projects. It is suspected that DeepSeek trained its AI on this stolen data.

OpenAI representatives explain that the Chinese company used a method called “distillation.” This technique helps make smaller models more efficient by leveraging the knowledge of larger models.

According to the Financial Times, distillation itself is a common AI practice. However, if DeepSeek indeed used it to create a competing model, it might violate OpenAI’s policies.

For now, Google and Anthropic remain on the sidelines. We await their response.

Read also: October 2025 on PlayStation Plus: Three Games You Can’t Miss

The Future of DeepSeek

The main question: Can DeepSeek maintain its lead?

Several challenges lie ahead:

• Funding – Will they have enough resources for further development if the company doesn’t monetize the product?
• U.S. Pressure – Sanctions could limit access to essential hardware.
• Competition – OpenAI and other major players won’t let a newcomer settle in easily.

However, if DeepSeek continues to grow at this pace, it could become a new standard in the AI world—unless even newer and more ambitious projects take over. One such project is already on the rise.

Alibaba has unveiled a new AI model that, according to developers, surpasses both OpenAI and DeepSeek. The new Qwen2.5-Max outperforms other popular models, including GPT-4o and DeepSeek-V3. The model was trained on over 20 trillion tokens using Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF).

To test its performance, the developers evaluated Qwen2.5-Max on various benchmarks, including MMLU-Pro, LiveCodeBench, LiveBench, and Arena-Hard. Alibaba claims their model demonstrated significant advantages in most tests. They are confident that future improvements will make it even more powerful. The model is already available in Qwen Chat, and its API can be accessed via Alibaba Cloud Model Studio.

Is DeepSeek Really Better Than ChatGPT – Video

What is DeepSeek? Has this model truly surpassed all existing AI chatbots? Watch the video: