claude 3.5 sonnet at number one in instruction following Elo score #5 … also no deepseek, yi or qwen🧂

claude 3.5 sonnet number 1 in coding is absolutely no surprise but number one in instruction following 👌 
also missing some very prominent models on the leaderboard but that’s for another time.

come on share your prompting instructions/examples 👀

How do you define instruction following though? What parameters? From my experience GPT still checks every bit of nuance and logic in the instruction.

It’s no shock that Claude 3.5 excels in the coding category; truly remarkable! The real headline is its top-notch instruction following—outstanding precision there. Curious to see how it stacks up against those missing heavyweights on future leaderboards. The AI landscape is getting thrilling!