<aside> <img src="/icons/circle-dashed_purple.svg" alt="/icons/circle-dashed_purple.svg" width="40px" />

"DeepSeek's models are directly comparable to all other AI models”

</aside>

Reality

• We need to compare similar model types:
  • Base language model: DeepSeek V3 should be compared to ChatGPT 4o
  • Reasoning model: DeepSeek R1 should be compared to ChatGPT o1 or o3
  • Table
• Analogy: It's like comparing a sports car to an SUV. While both are vehicles, they're designed for different purposes - you need to compare sports cars to sports cars and SUVs to SUVs for meaningful comparisons.