- DeepSeek R2 promises up to 97% lower processing costs than GPT-4o.
- The model uses 1.200 billion parameters and has been trained on 5,2 petabytes of data.
- 82% of the chips used in your training are Huawei Ascend, reducing dependence on NVIDIA.
- Its efficiency and price could make it a serious global competitor.
The sector of Artificial Intelligence is experiencing a true technological race, where large companies fight to obtain more powerful and efficient models. DeepSeek, a Chinese-based company, has burst onto the scene with its next release: DeepSeek R2, which according to leaked data, could be a game changer by offering a IA high capacity with costs considerably lower than those of its most direct rivals.
The news has spread like wildfire in the technology sector, as DeepSeek R2 aims to become a viable and affordable alternative to giants like ChatGPT, Google Gemini or Meta Llama. The desire to compete head-to-head is based on a compelling argument: The cost of processing large volumes of information will, according to estimates, be much lower than that of currently cutting-edge models.
DeepSeek R2 Technical Features
DeepSeek R2 is presented as an open-source, generative artificial intelligence model, allowing both companies and individual developers to research, customize, and adapt the technology to their needs. Its new version implements 1,2 trillion parameters, thus raising the bar compared to its predecessor and approaching the performance of the most advanced options on the market.
The model has been trained with the help of 5,2 petabytes of data, Most of them come from the C-Eval 2.0 suite, a detail that highlights the magnitude of the work performed. In terms of capabilities, its performance in computer vision is expected to achieve outstanding results, reaching up to 92,4% accuracy in tests with the COCO system.
Cost of use is one of its most important advantages. According to the leaks, operating with DeepSeek R2 will cost just $0,07 per million tokens input and $0,27 per million tokens output, representing a cost reduction of nearly 97% compared to GPT-4.
Technological independence with Huawei Ascend chips
One of the most relevant changes in the development of DeepSeek R2 is the commitment to hardware national. 82% of the chips used during training correspond to the Huawei Ascend 910B series, marking a clear distance from the use of NVIDIA GPUs, something common in the sector until now. This decision involves not only a drastic reduction in energy and production costs, but also greater technological sovereignty, especially for the Chinese market.
This shift toward homegrown solutions allows DeepSeek to continue to move forward independently and reduce its dependence on the US supply chain, which could signal a shift in future AI development in both China and other emerging markets.
A global competitor against AI titans
In a landscape where innovation is constant, DeepSeek R2 is emerging as a rival that could challenge popular models like the GPT-4 Turbo or Google Gemini 2.0 Pro, thanks to its price-performance ratio. The focus on efficiency and scalability, along with its open-source philosophy, makes it attractive to both businesses and independent developers looking for more affordable alternatives without sacrificing performance.
For now, most of the available details come from leaks and analysis by AI experts, and we'll have to wait for official confirmation from the company. However, the excitement generated by this new model is undeniable and marks a turning point in the cost trend for generative artificial intelligence.
The trend toward increased competition in the field of AI continues, driving advances that democratize access to these technologies. DeepSeek R2 represents a notable step in the search for more accessible and sustainable solutions, capable of democratizing access to advanced artificial intelligence both within and outside Asia.
Passionate writer about the world of bytes and technology in general. I love sharing my knowledge through writing, and that's what I'll do on this blog, show you all the most interesting things about gadgets, software, hardware, tech trends, and more. My goal is to help you navigate the digital world in a simple and entertaining way.