The DeepSeek frenzy is reshaping the market for large language models (LLM). In addition to open-source and closed-source models, the open-closed-source composite (hybrid) model offers another direction of development.
The open-source model is transparent, but characterized by a low return on investment (ROI) and slow iterations. The closed-source model is flexible in fine-tuning but lacks transparency.
The hybrid model combines the characteristics of both. Google and xAI have already turned their direction towards the hybrid model, which is expected to become a new trend in LLM development.
Open-source LLM development started earlier than closed-source
Chart 1: Development timeline of key open- and closed-source LLMs, 2019-2025
Low ROI and slow iterations are obstacles for open-source LLMs
Low transparency and lack of fine-tuning impact closed-source LLMs
HybridLLMs strike a balance between tech sharing and commercial sustainability
Chart 3: Applicable scenarios and payment methods of key LLMs
DeepSeek's medium-term goal is to reshape the global LLM market
Chart 4: DeepSeek short-to-long-term goals with open-source model
