humaneval
A pip-installable benchmark runner for LLMs and agents. Five minutes to your first eval.
A strict, auditable HumanEval benchmark runner for GGUF models served via llama.cpp, using its OpenAI-compatible HTTP API.
