The release of the Chinese AI app DeepSeek has sent shockwaves through financial markets and the global tech industry. DeepSeek’s claim that its R1 artificial intelligence model was developed at a fraction of the cost of its competitors has not only challenged industry norms but has also led to significant repercussions for some of the world’s largest tech companies. In a short period, DeepSeek has risen to become the most downloaded free app in the US, a feat that has brought it into direct comparison with established AI giants like OpenAI’s ChatGPT and Google’s Gemini.
This comparison highlights a critical moment in the AI industry, where innovation is not exclusive to the traditional powerhouses of Silicon Valley. With a mix of unique features and cost advantages, DeepSeek is forcing its competitors to take notice. But how does it actually stack up against the widely recognized ChatGPT in terms of capabilities?
Writing Assistance
One of the most popular uses of AI chatbots is writing assistance. ChatGPT has long been celebrated for its ability to generate content ranging from summaries to detailed blog posts, making it a valuable companion across workplaces.
To test both platforms, a task was set: summarize the best Scottish football players of all time and draft a blog post about them. DeepSeek responded quickly with a top-ten list, placing Kenny Dalglish, the legendary Liverpool and Celtic player, at number one. Its response included details about each player’s position, clubs, and notable achievements. Surprisingly, DeepSeek also included two non-Scottish players—Brian Laudrup and Henrik Larsson—explaining their impact on Scottish football and acknowledging Larsson’s Swedish nationality.
ChatGPT’s response was equally detailed, offering a similar list of names with additional context about their careers. It concluded with a statement emphasizing the timeless legacy of these players. Both platforms demonstrated competence, but their approaches differed. DeepSeek excelled in providing structured, concise answers, while ChatGPT leaned toward delivering more detailed narratives.
Coding Capabilities
The advent of advanced AI models has been transformative for programmers, offering assistance in everything from debugging to complex problem-solving. ChatGPT has been a go-to tool for developers, with its occasional outages prompting amusing complaints from users on social media about their reliance on the chatbot.
DeepSeek, however, appears to be a serious contender in this space. Javier Aguirre, an AI researcher at Samsung Medical Center in South Korea, recently shared his experience with DeepSeek, praising its ability to solve a particularly tricky coding issue that ChatGPT’s o1 model could not resolve. His experience demonstrates that DeepSeek’s reasoning capabilities in coding may exceed expectations.
Notably, some professionals are combining multiple models to achieve optimal results. Addy Osmani, Google’s Head of Chrome Developer Experience, mentioned using DeepSeek in tandem with Anthropic’s Claude Sonnet for coding. Osmani highlighted DeepSeek’s cost advantages, describing it as significantly cheaper than both Claude Sonnet and OpenAI’s o1 model. This affordability, coupled with impressive performance, gives DeepSeek an edge in the programming arena.
