In just a week since its launch, DeepSeek, a Chinese competitor to ChatGPT , has exploded in popularity and sent the American tech industry into a frenzy. The new chatbot is on par with AI market leaders and is offering its services in an open-source model .
DeepSeek – a challenge for OpenAI and the entire industry
Since the introduction of ChatGPT by OpenAI, it seemed that the American AI leader had no equal. However, the launch of DeepSeek-R1 changed the balance of power. The Chinese model, based on the efficient use of technology, proved that it is possible to create solutions comparable to the most advanced systems on the market at a fraction of the cost .
DeepSeek-V3 , on which the R1 model is based, cost just $5.6 million to train – an almost laughable amount compared to the hundreds of millions OpenAI spends on its models.
DeepSeek offers its services on an open-source basis , meaning that any user can modify and customize the model to philippines email list suit their needs. This is a key advantage over ChatGPT, which remains a closed system that requires paid subscriptions.
Technology that surprises with its possibilities
DeepSeek-R1 operates on a 671 billion-parameter model and is considered one of the most efficient AI systems. In meta tests ads on threads – platform monetization benchmark tests, it achieves results at a similar level to the latest OpenAI models , surpassing them in some applications. However, the key advantage is the price – access to the DeepSeek AI revolution API costs just $2.19 per million tokens, while in the case of OpenAI this cost is a whopping $60.
The Chinese startup also uses technology
-machine-with-application-examples” target=”_blank” rel=”noreferrer noopener nofollow”>reinforcement school email list learning , which allows the model to better “understand” thought processes and create more accurate answers. Additionally, DeepSeek can supplement its data from Internet search engines in real time , which increases its functionality.
The American Response – Stargate Project and OpenAI’s Free Model
DeepSeek forced OpenAI to react quickly. In response to the Chinese launch, Sam Altman, OpenAI’s CEO, announced that the new o3-mini model would be available for free with a limited number of prompts , an attempt to attract users who had begun to drift away from the Chinese competitor.
At the same time, the US administration is focusing on developing its own AI technologies as part of the $500 billion Stargate project . This project is intended to help American companies combat the growing dominance of Chinese technology companies.