ABench is an evolving open-source benchmark suite designed to rigorously evaluate and enhance Large Language Models (LLMs) on complex cross-domain tasks.
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).