From GPT-4 to Mistral 7B, there is a 300x range in the cost of LLM inference(twitter.com)2 points by Gcam 2 years ago | 0 commentsNo comments yet