Artificial analysis’ leaderboard for models is somewhat useful for comparing model intelligence, but does not take into account quantization for open models. Is there a way to better compare quantized open models against each other and proprietary models other than running them directly
