How well can LLMs actually build transactions on Solana? Until now, there wasn’t a simple, reproducible way to measure. Introducing Solana Bench — open benchmarks for testing AI’s operational competence on Solana. Mega-thread 🧵