How DeepSeek’s budget-friendly AI model stacks up against ChatGPT, Claude, and Gemini in SQL, EDA, and machine learning
Nvidia stock price slumped over 15% on Monday, Jan 27th, after a Chinese startup, DeepSeek, released its new AI model. The model performance is on par with ChatGPT, Llama, and Claude but at a fraction of the cost. According to Wired, OpenAI spent more than USD$100m to train GPT-4. But DeepSeek’s V3 model was trained for just $5.6m. This cost efficiency is also reflected in the API costs — for every 1M tokens, the deepseek-chat model (V3) costs $0.14, and the deepseek-reasoner model (R1) costs only $0.55 (DeepSeek API Pricing). Meanwhile, gpt-4o API costs $2.50 / 1M input tokens, and o1 API costs $15.00 / 1M input tokens (OpenAI API Pricing).
Always intrigued by emerging LLMs and their application in data science, I decided to put DeepSeek to the test. My goal was to see how well its chatbot (V3) model could assist or even replace data scientists in their daily tasks. I used the same criteria from my previous article series, where I evaluated the performance of ChatGPT-4o vs. Claude 3.5 Sonnet vs. Gemini Advanced on SQL queries, Exploratory Data Analysis (EDA), and Machine Learning (ML).
