What Enterprise Is Getting For Their LLM Dollars
SMRTR summary
The cost of large language models (LLMs) is dropping rapidly, with performance improving 10x annually. This trend is driven by smaller models, open-source options, and optimizations. GPT-3 achieved an MMLU score of 42 at $60 per million tokens, but costs have since decreased tenfold. This price reduction is making AI more accessible and enabling new enterprise use cases. However, the market remains unpredictable, with experts coining terms like "LLMflation" to describe the decreasing cost of AI tokens.
SMRTR provides this summary for quick context. The original article belongs to Forbes.
Read the original article