DeepSeek vs OpenAI – Comparison Breakdown and What It Means For Your Tech Startup

DeepSeek has been making headlines—and for good reason. Their recently released R1 model topped benchmarks, only to be surpassed just a few days later by OpenAI’s o3-mini. In that same time span, DeepSeek also released a multimodal model, Janus-Pro-7B, which outperformed DALL-E 3 on a variety of benchmarks. The AI races is on.

In this session, we’ll cover: – How DeepSeek’s R1 and Janus-Pro-7B models diverge from OpenAI’s o1, o3-mini and DALL-E 3 in architecture and training methodology. – A side-by-side look at how these models stack up in performance, benchmark tests, and real-world prompt evaluations. – The impact of reinforcement learning, chain-of-thought reasoning, and prompt engineering for reasoning models.

Speakers: – Dan Cleary – Co-founder of PromptHub

– Tony Yang – Head of Growth at Mucker Capital