However, if we were to support BYOK on every feature in Tana, we would not be able to apply multiple models for different tasks and make use of the rapid advances that different players are driving. The fine-tuning of Tana models increases the quality of the results of LLMs, both when it comes to reducing hallucinations and working with the knowledge graph and Tana concepts. Additionally, our integrated AI gives us a total volume that provides better prices through partnerships with OpenAI, Anthropic, and other LLM companies than one user could obtain, thus making it cheaper for each user.