The growing adoption of open source AI models is transforming the AI landscape, but frontier model providers like Anthropic remain relevant due to their dominance in early-stage deployments and premium pricing.
The AI Life Cycle: Frontier Models and Open Source Alternatives
Decagon CEO Jesse Zhang recently proposed a new perspective on the relationship between frontier and open source AI models, suggesting they're not competitors, but rather two phases of the same life cycle. As AI deployments mature, they switch to lighter, more cost-effective open source models, while new use cases continue to emerge, driving demand for frontier models.
This two-tiered economy is evident in data from Vercel's AI gateway dashboard and OpenRouter, which show that while open source models like DeepSeek are gaining traction in terms of token volumes, frontier models like Anthropic's continue to dominate overall spend due to their premium pricing.
Usage Trends and Pricing: A Closer Look
Some key statistics illustrate this trend:
- DeepSeek processed over a third of the tokens on Vercel's infrastructure in the past week.
- Anthropic still accounts for more than half of the overall AI spend on Vercel.
- OpenRouter reports that DeepSeek V4 Flash processes 5.3 trillion tokens weekly, while the frontier model Opus 4.8 handles just over 2 trillion.
- The average token cost for Opus 4.8 is roughly 23 times higher than V4 Flash.
The market's growth and the complexity of certain use cases are likely contributing factors to the continued relevance of frontier models. As Zhang puts it, "The frontier labs will keep owning discovery. Open source will increasingly own production."
What's Next for the AI Economy?
As the AI landscape continues to evolve, we can expect to see new developments, such as Nvidia's Nemotron model, which is poised to become a major player due to its adaptability and Nvidia's strong connections. The stable coexistence of frontier and open source models may become a defining feature of the AI economy.
